sign in sign up
<<
Home , Bush moa, Cominella, Flax weevil, Fluttering shearwater, Heiji Rebellion, Latridopsis

Tuhinga: Records of the Museum of Te Papa Tongarewa

The journal of scholarship and mätauranga

Number 29, 2018 Tuhinga: Records of the Museum of New Zealand Te Papa Tongarewa is a peer-reviewed publication, published annually by Te Papa PO Box 467, , New Zealand

TE PAPA® is the trademark of the Museum of New Zealand Te Papa Tongarewa

Tuhinga is available online at www.tepapa.govt.nz/tuhinga

It supersedes the following publications: Museum of New Zealand Records (1171-6908); National Museum of New Zealand Records (0110-943X); Dominion Museum Records; Dominion Museum Records in Ethnology.

Editorial board: Puawai Cairns, Sarah Farrar, Bronwyn Labrum, Sean Mallon, Claire Regnault, Kirstie Ross, Mark Stocker, Susan Waugh

ISSN 1173-4337

All papers © Museum of New Zealand Te Papa Tongarewa 2018 Published June 2018

For permission to reproduce any part of this issue, please contact the editorial co-ordinator, Tuhinga, PO Box 467, Wellington.

Edited by Mark Stocker Cover design by Tim Hansen Design by Christine Lin Barraud, Big Catch Creative Typesetting by Susi Bailey Tuhinga: Records of the Museum of New Zealand Te Papa Tongarewa Number 29, 2018

Contents A new of Pliocene (Aves: ) from New Zealand 1 Alan J.D. Tennyson and Al A. Mannering

New locality records for two species of protected , fairburni 20 (Brookes, 1932) and Hadramphus stilbocarpae Kuschel, 1971 (Coleoptera: ), from southern Fiordland, New Zealand Colin M. Miskelly, Alan J.D. Tennyson and Colin R. Bishop

Alexander McKay: New Zealand’s first scientific photographer 35 Simon Nathan

Rising sun at Te Papa: the Heriot collection of Japanese art 50 David Bell and Mark Stocker

Characterisation of Fifteen Sources of Japanese Obsidian: PIXE-PIGME 77 analysis, and identification of archaeological specimens B.F. Leach, S. Oda and J.R.

In search of the Archaic: Archaeological excavations at 90 Sarah’s Gully, Coromandel Peninsula, New Zealand from 1956 to 1960 Janet Davidson

Increasing visitor engagement in a contemporary art exhibit through 165 a participatory activity Edith MacDonald

Evaluation of Contraception: Uncovering the collection of Dame Margaret Sparrow 174 Stephanie Gibson Tuhinga 29: 1–19 Copyright © Museum of New Zealand Te Papa Tongarewa (2018) 1

A new species of Pliocene shearwater (Aves: Procellariidae) from New Zealand Alan J.D. Tennyson* and Al A. Mannering**

* Museum of New Zealand Te Papa Tongarewa, PO Box 467, Wellington 6140, New Zealand ([email protected]) ** 8 Roswell Place, 8042, New Zealand ([email protected])

ABSTRACT: We describe two partial but well-preserved Late Pliocene skeletons from , New Zealand, as a new species of . In structure, these bones match those of a shearwater (: Procellariidae) but the new is distinguished from all known extant and extinct taxa by a unique combination of features. It was a gliding species as large as the largest species of extant shearwater. It represents the first pre- Pleistocene record of a new shearwater taxon from the western Pacific and helps reveal the history of shearwater evolution. Today, New Zealand has the greatest diversity of breeding shearwater species in the world, and the new fossil adds weight to other evidence that have a long history in this region.

KEYWORDS: Procellariiformes, Procellariidae, shearwater, fossil, Pliocene, biogeography, New Zealand.

Introduction shearwaters are sister taxa, and the other large shearwaters are sister to that (Austin 1996; Heidrich et al. 1998; Shearwaters are among the most diverse group of Nunn & Stanley 1998; Austin et al. 2004; Penhallurick & Procellariiformes (, petrels, prions and shearwaters). Wink 2004; Onley & Scofield 2007). Phylogenetic studies have found the shearwaters to be a The of the shearwaters (Procellariidae) monophyletic clade within the Procellariidae (e.g. is controversial but we follow Dickinson and Remsen Heidrich et al. 1998; Nunn & Stanley 1998; Penhallurick & (2013), who recognised three genera: Calonectris for Wink 2004). Molecular studies suggest that the closest relative the streaked shearwater (C. leucomelas) and the Scopoli’s of the shearwater clade is the Kerguelen petrel Lugensa shearwater (C. diomedea clade); Ardenna for all the other brevirostris, then successively more distantly the petrels larger taxa – wedge-tailed shearwater (A. pacifica), Procellaria, Bulweria and Pseudobulweria, with other procellariid Buller’s shearwater (A. bulleri), pink-footed shearwater genera (prions Pachyptila, blue petrel Halobaena caerulea, the (A. creatopus), flesh-footed shearwater (A. carneipes), great fulmars and the gadfly Pterodroma petrels) being even more shearwater (A. gravis), sooty shearwater (A. grisea) and distant (Nunn & Stanley 1998; Penhallurick & Wink 2004). short-tailed shearwater (A. tenuirostris); and for Within the shearwaters, there are three main : the smaller taxa. Kuroda’s (1954) pioneering work on Calonectris (470–1060 g); all the other large shearwaters shearwater relationships recognised Calonectris as distinct (320–950 g); and all the smaller (120–575 g) shearwaters and placed all other species in the Puffinus, which (the Manx-group and little/Audubon’s complex, as defined he then divided into subgenera: P. pacificus and P. bulleri in by Onley & Scofield 2007). Calonectris and all the smaller Thyellodroma; P. creatopus and P. carneipes in Hemipuffinus; 2 Tuhinga, Number 29 (2018)

P. gravis and the extinct P. conradi in Ardenna; P. griseus, pelagornithid taxa, possibly Pelagornis miocaenus from the P. tenuirostris and the Christmas shearwater (P. nativitatis) Middle to Late (Scarlett 1972; Harrison & Walker in ‘Puffinus Neonectris’; and all the smaller taxa in ‘Puffinus 1976) and Neodontornis stirtoni from the Puffinus’. Using a molecular phylogenetic analysis, Austin to Pliocene (Howard & Warter 1969); another Pliocene (1996) concluded that the Puffinus taxa of Kuroda fell into pelagornithid of unclear affinities (McKee 1985); and a two distinct clades: all the larger taxa and all the smaller mid-Pliocene shearwater skull, also of uncertain affinities taxa (including P. nativitatis, which had previously been (Henderson & Gill 2010). Additionally, a tarsometatarsus linked with P. griseus and P. tenuirostris). Using further DNA from the Late or Early Paleocene may be from a analysis, Heidrich et al. (1998) concluded that the two seabird (Ksepka & Cracraft 2008). clades identified by Austin were not sister taxa, therefore The few described from the Pliocene of supporting the conclusion that all three shearwater clades Taranaki, New Zealand, are the pelagornithid of unclear should be afforded generic rank. affinities mentioned above and a penguin (McKee 1988), Olson and Rasmussen (2001: 254) noted that ‘the species although mention has been made of several others being of shearwaters are marked by a progression from a primitive, found, including shearwaters (see McKee 1994; Henderson aerially adapted condition (Calonectris) to increasing use & Gill 2010; Worthy & Tennyson 2010: 331). of both the wings and feet for underwater propulsion … Many fossil species of procellariiform have been named in which the becomes flattened, the forewing and the most speciose clade of extinct taxa is the shearwaters shortened … the femur stouter and more curved … and the (Warham 1996; Warheit 2002). Pre-Pleistocene shearwater tarsometatarsus more laterally compressed’. They considered are widely distributed, but are found mainly in that this progression was from Calonectris to the group Miocene and Pliocene North Atlantic and North Pacific containing Ardenna pacifica, A. bulleri, A. creatopus and A. coastal deposits (Warheit 2002). Shearwater fossils of this age carneipes, then to A. gravis, then to the Puffinus taxa. are surprisingly rare in the southern hemisphere, consisting Although New Zealand today is the centre of diversity for of a range of modern or unidentified taxa from South shearwaters, with nine (of a worldwide total of 42) breeding Africa (Olson 1985a,b), the western South American taxa (Dickinson & Remsen 2013), the pre-Late Pleistocene (Hoffmeister et al. 2014) and the various unidentified fossils fossil record of Procellariiformes, including shearwaters, from New Zealand (see above). The earliest fossil species in the entire western Pacific is minimal (Warheit 2002; of shearwater globally is Puffinus raemdonckii (van Beneden, Henderson & Gill 2010; Worthy & Tennyson 2010). The 1871) from the Early of Belgium, however its discovery of pre-Late Pleistocene fossil procellariiform true affinities are uncertain (van Beneden 1871; Brodkorb specimens in New Zealand can, therefore, add significant 1962; Olson 1985c; Mayr & Smith 2012). Several fossil new knowledge about the history of this group. species of shearwater are known from the Miocene through New Zealand has one of the best fossil records worldwide to the Holocene, with the earliest of these being Puffinus for for the Late Pleistocene–Holocene (Worthy & micraulax Brodkorb, 1963, from the Early Miocene of South Holdaway 2002). However, the pre-Late Pleistocene avian Carolina, USA (Brodkorb 1963a). fossil record is poor, apart from penguin fossils and those Here we report on significant new shearwater fossils from from an Early Miocene lacustrine site in Otago (e.g. Simpson the Pliocene of Taranaki, and compare them with all living 1975; Fleming 1979: 75; Fordyce 1991; Worthy et al. 2007; and extinct shearwater taxa. Henderson & Gill 2010; Worthy & Tennyson 2010; Ksepka et al. 2012; Mayr et al. 2017). Methods The few pre-Late Pleistocene non-penguin New Zealand Dave Allen collected one of the fossils in an Upper Pliocene seabirds reported include: the Paleocene Australornis lovei concretion at Ohawe Beach, south Taranaki, in 2006. The described by Mayr & Scofield (2014); a Paleocene tropicbird specimen was found by splitting a mudstone concretion in described by Mayr & Scofield (2015); a fragmentary furcula half with a sledgehammer. Unfortunately, the counterpart of uncertain relationships named Manu antiquus from the of the concretion was not collected and it is unknown if Middle to Late Oligocene (see Worthy & Tennyson 2010); there were further bones in it. Allen and Craig (then of the Early Miocene diving petrel Pelecanoides miokuaka the Institute of Geological and Nuclear Sciences) roughly described by Worthy et al. (2007); an indeterminate prepared the specimen, with final preparation work by Al diving petrel (Scofield et al. 2006); two Mannering. A new species of Pliocene shearwater (Aves: Procellariidae) from New Zealand 3

A second specimen was found in 2011 by Xavier recognisable elements – one vertebra; pL humerus; most Johnson, by splitting a concretion at Waihi Beach, about 1 of the R humerus, although the shaft is damaged near the km southeast of the first specimen, in similarly aged rocks proximal end so an accurate length measurement could not be (e.g. McKee 1985, 1988, 1994). Initial preparation was obtained; complete L ulna; most of the R ulna; most of both carried out by Alastair Johnson and it was completed by Al radii; complete L carpometacarpus; dR carpometacarpus; Mannering. complete L manus phalanx p digitus major; complete L The relevant morphological characters of the new femur; pR femur; pL tarsometatarsus; dR tarsometatarsus – Taranaki specimens were compared with representatives and four unidentified small fragments. Most bones are partially of all procellariiform families and all key species within obscured because of the way they are lying or because other Procellariidae. Key references containing morphological bones lie on top of them, and early preparation involved a characters were Chandler (1990), Kuroda (1953, 1954, roughly applied consolidant, which obscured some bone 1955), Olson (2008, 2009) and Olson & Rasmussen (2001). features and could not be removed by Mannering. The Important species of procellariid not directly compared were specimen lacks the skull, all but one vertebrae, all pectoral Scopoli’s shearwater (Calonectris diomedea) and Cape Verde and pelvic elements, the outermost wing bones, both shearwater (C. edwardsii), which were not available to us tibiotarsi and all but one of the phalanges. The specimen is but are considered to be closely related to Cory’s shearwater osteologically mature, showing the degree of bone fusion (C. borealis) (e.g. Olson 2008). We did not compare every expected in an adult. species in the small-bodied shearwater clades because initial Humerus (proximal end) with caput humeri undercut comparisons showed little similarity between these groups and overhanging the fossa pneumotricipitalis dorsalis slightly; and the Taranaki fossil. Also, comparison was made with capital shaft ridge extends to caput humeri; no tubercle on the extinct Mediterranean genus and species the cranial surface of the ventral margin of the bicipital area; of the procellariid Pterodromoides minoricensis based on the crista deltopectoralis rounded in dorsal profile. description by Seguí et al. (2001). Humerus shaft with rounded cross section (rather than Osteological terms are based on the work of Baumel and dorsoventrally flattened), with ratio of central shaft width Witmer (1993) and Livezey and Zusi (2006); p = proximal, to depth of 1.3. Distal end of the humerus, processus d = distal, L = left, R = right. Measurements were taken supracondylaris dorsalis (= ectepicondylar process) in millimetres (mm) using Vernier callipers. The primary moderately sized; fossa m. brachialis medium depth. Ulna comparative specimens used are listed in the Appendices. All (proximal end), caudal surface of the shaft forms a prominent are part of the Museum of New Zealand Te Papa Tongarewa and continuous ridge to the tip of the proximal end. Femur (Te Papa) collection (NMNZ, including specimens with ‘S’ shaft in medial view fairly curved (arched dorsally). and ‘OR’ prefixes). Systematic Palaeontology Holotype locality and age: Ohawe Beach, south Taranaki, New Zealand (39˚35.72′ S, 174˚12.90′ E); Waipipian (3.0– PROCELLARIIFORMES 3.6 Ma, Late Pliocene). (FR Number Q21/F0175.) Mollusc Family PROCELLARIIDAE fossils in the concretion were identified as Dosinia sp. cf. D. Leach, 1820 lambata and Maorimactra sp. cf. M. ordinaria (B. Marshall, pers. comm., 2014). Dosinia lambata has a currently accepted age as Ardenna Reichenbach, 1853 ranging from Duntroonian (Late Oligocene) to Holocene, and Type Species: Ardenna gravis (O’Reilly, 1818) M. ordinaria is also extant and its fossils are found as far back Ardenna davealleni Tennyson & Mannering, sp. nov as the Opoitian (Pliocene) (B. Marshall, pers. comm., 2014). These age ranges are consistent with the Waipipian (Pliocene) Holotype: NMNZ S.45183: Partial skeleton (Fig. 1), age previously attributed to these deposits (e.g. McKee 1985, collected by D. Allen, 8 October 2006. 1988, 1994). Minimum storm base wave depth indicates that Description of holotype: See Fig. 1 and Appendices 1 and 3. the fossil bird’s depositional water depth may have been 50 m The specimen is a well-preserved individual, which was or more (Peters & Loss 2012) in order for the skeleton to have fossilised with some disturbance to the skeleton before burial, remained partly intact. However, the presence of some of the such that several key bones have been lost. It preserves 14 mollusc species in ‘life position’ suggests a shallower depth, as 4 Tuhinga, Number 29 (2018)

Fig. 1 Pliocene fossil shearwater Ardenna davealleni n. sp. holotype NMNZ S.45183. A, general view of the entire fossil: 1, right humerus; 2a, proximal end of the left humerus; 2b, distal end of the left humerus; 3a, distal end of the right tibiotarsus; 3b, ?Proximal end of the right tibiotarsus; 4, unidentified fragment; 5, left ulna; 6, left radius; 7, proximal end of the right femur; 8, left carpometacarpus; 9, distal end of the right carpometacarpus; 10, left femur; 11, unidentified fragment; 12, proximal end of the left tarsometatarsus; 13, right ulna, missing the distal end; 14, right radius; 15, left manus phalanx proximal digitus major; 16, vertebra. Scale bar is 5 cm. B, detailed view of the proximal end of the left humerus, showing Appendix 2 characters 1, 2, 3 and 4; character 3 is on the cranial surface of the bone at the arrowed point. C, detailed view of the distal end of the right humerus, showing Appendix 2 characters 6 and 7. D, detailed view of the proximal end of the right ulna, showing Appendix 2 character 8. E, detailed view of left femur, showing Appendix 2 character 9. Scale bar in B–E is 1 cm. A new species of Pliocene shearwater (Aves: Procellariidae) from New Zealand 5

Fig. 2 Pliocene fossil shearwater Ardenna davealleni n. sp. paratype NMNZ S.46316. 1, left carpometacarpus; 2a, proximal end of the right ulna; 2b, distal end of the right ulna; 3a, cnemial crest of the proximal end of the right tibiotarsus; 3b, proximal end of the right tibiotarsus; 3c, distal end of the right tibiotarsus; 4, right fibula; 5, proximal end of the right carpometacarpus; 6a, proximal end of the right humerus; 6b, distal end of the right humerus; 7a, proximal end of the right radius; 7b, shaft of right radius; 8, left tarsometatarsus; 9a, proximal end of the left humerus; 9b, distal end of the left humerus; 10a, proximal end of the left radius; 10b, distal end of the left radius; 11a, proximal end of the left ulna; 11b, distal end of the left ulna. 6 Tuhinga, Number 29 (2018)

specimens in the Te Papa collection (n = 185 ‘lots’ for both processus supracondylaris dorsalis, a larger humerus to ulna mollusc species combined) were collected live between 4 m length ratio, a smaller ulna to carpometacarpus length ratio, and 48 m (B. Marshall, pers. comm., 2015; see also McKee a smaller ulna to manus phalanx p digitus major length ratio, 1994). and a longer cnemial crest. Etymology: The species name honours Dave Allen of New The fossil species differs from all other species in the genus Plymouth, the collector and donor of the holotype. Ardenna by having a smaller ulna to manus phalanx p digitus major length ratio. It differs from A. grisea and A. tenuirostris Suggested English vernacular name: Pom’s shearwater, as in having a rounded humeral shaft, a deeper fossa brachialis, requested by Dave ‘Pom’ Allen. a relatively straighter femur shaft, a relatively shorter cnemial Paratype: NMNZ S.46316: Partial skeleton (Fig. 2), collected crest, and a less laterally flattened tarsometatarsus. The by X. Johnson, 6 June 2011. relative length of the cnemial crest is intermediate between that found in most species of Ardenna and that of A. grisea and Description of paratype: See Fig. 2 and Appendices 1, 3 and 4. A. tenuirostris. The species differs from A. carneipes, A. This specimen is also a well-preserved individual, with creatopus and A. gravis in having a more rounded crista 11 recognisable (mostly complete) elements: L humerus, deltopectoralis, and a smaller processus supracondylaris missing the very proximal end; R humerus; both ulnae; dorsalis. It differs from A. pacifica, A. bulleri, A. carneipes both radii; L carpometacarpus, missing the proximal end; R and A. creatopus in having a higher humerus to ulna carpometacarpus, with the distal end still embedded in matrix; length ratio. It differs from A. pacifica, A. bulleri, A. R tibiotarsus; R fibula; and R tarsometatarsus. The specimen carneipes, A. creatopus and A. gravis in having lower ulna is osteologically mature and has the same morphological to carpometacarpus and ulna to femur length ratios, and features described for the humeri and ulnae of the holotype. in having a relatively longer cnemial crest. It differs from The relatively short cnemial crest of the tibiotarsus and A. bulleri in having a less undercut caput humeri. It shares the relatively laterally broad tarsometatarsus are consistent most characters with A. bulleri, but it is substantially larger with that seen in some other species of Ardenna. As for the holotype, the partially intact skeleton indicates that the bird’s than that species and is larger than all members of the genus remains were originally deposited in water 50 m deep or except A. gravis. Its wing and leg bones are in the range of more, although deposition was possibly shallower than this 14–24% longer than those of A. bulleri. It differs from (see above). This skeleton is from a bird very similar in size fossil taxa of Ardenna primarily by a combination of size, to the holotype. proportions and age.

Paratype locality and age: Waihi Beach, south Taranaki, New Zealand (39˚36.10′ S, 174˚14.05′ E); Waipipian (3.0– Further comparisons 3.6 Ma, Late Pliocene; McKee 1985, 1988, 1994). with living Measurements and ratios of holotype and paratype procellariiform taxa bones: See Appendices 3 and 4. See Figs 1 and 2 and Appendices 1, 3 and 4. Diagnosis Using an analysis of some key characters, all procellariiform families and all other procellariid genera can be differentiated The fossils are referred to the genus Ardenna. Compared from all shearwaters by a range of morphological features (see with species of Puffinus, the humerus of the fossil species Appendix 1). For example, shearwaters differ from all other has a rounded shaft (rather than dorsoventrally flattened), a procellariid genera by having a shallower fossa brachialis deeper fossa brachialis and a less proximally angled processus on the humerus. Aside from Daption, Pagodroma, Bulweria supracondylaris dorsalis, a larger ulna to carpometacarpus and Pseudobulweria, all other procellariid genera differ from length ratio, a straighter femur shaft, and a relatively shorter shearwaters by having a raised tubercle on the cranial surface cnemial crest on the tibiotarsus. Compared with species of the ventral margin of the bicipital area of the humerus. of Calonectris, the humerus of the fossil species has a more Additionally, fulmars do not have such ventrally prominent rounded profile to the crista deltopectoralis and a smaller crista medialis hypotarsi on the tarsometatarsus and, while A new species of Pliocene shearwater (Aves: Procellariidae) from New Zealand 7 shearwaters and many other procellariiforms have ventrally Detailed comparisons with fused cristae hypotarsi, other genera – such as Thalassoica, Lugensa and Pterodroma – show no such fusion, with fossil shearwater taxa albatrosses having the most extreme condition, whereby the For fossil shearwater taxa, Olson (2009: 468) concluded that cristae hypotarsi remain widely separated. Furthermore, only Puffinus micraulax Brodkorb, 1963 (Early Miocene, South some shearwaters and Pagodroma have a similarly rounded Carolina, USA), P. inceptor Wetmore, 1930 (Middle Miocene, angle (point) to the crista deltopectoralis on the humerus as California, USA), P. calhouni Howard, 1968 (Late Miocene, found in Ardenna daveallenii, and A. davealleni differs from California), P. barnesi Howard, 1978 (Late Miocene, procellariid genera (other than some shearwaters) by having California), P. felthami Howard, 1949 (Early Pliocene, a fairly curved femur shaft. California), P. kanakoffi Howard, 1949 (Late Pliocene, Ardenna davealleni is one of the largest species of California), P. priscus Miller, 1961 (Late Miocene, California) shearwater known. Based on A. davealleni having a femur of overall similar size to that of the gliding and P. mitchelli Miller, 1961 (Late Miocene, California) all Calonectris borealis, we estimate that the species had a belonged to ‘the more derived species of the genus Puffinus’ similar weight to the latter (see Campbell & Marcus owing to their relatively flattened humerus. Using the 1992), which is about 605–1060 g (see Marchant & taxonomy adopted in this paper, Olson’s definition would Higgins 1990; Onley & Scofield 2007). In comparison, include species of Puffinus, together with Ardenna grisea and A. bulleri weighs about 385–490 g (Marchant & Higgins A. tenuirostris – all clearly separable from A. davealleni, which 1990; Onley & Scofield 2007), which is only half the weight has a rounded humeral shaft (see Appendix 1). The humerus of C. borealis. It should be noted that shearwaters exhibit of P. diatomicus Miller, 1925 (Late Miocene, California) has only very limited sexual size dimorphism, with males being a ‘flat shaft’ also (see Miller 1925). Puffinus tedfordi Howard, larger on average (Marchant & Higgins 1990; Warham 1971 (Early Pliocene, Mexico) belongs to this diving group 1990). also based on its laterally flattened tarsometatarsus (Olson Within the shearwaters, Puffinus species differ from 2009). Miller (1961) and Chandler (1990: table 7) detailed species of Calonectris and most species of Ardenna in having the high degree of flattening in the humeri of P. inceptor, P. a dorsoventrally flatter humeral shaft and a shallower fossa calhouni, P. barnesi, P. felthami, P. kanakoffi, P. priscus and P. brachialis on the humerus. Puffinus species differ from other mitchelli. Miller (1961) noted additionally that P. inceptor had shearwaters in having a smaller processus supracondylaris features on the distal end of the humerus – the distal position dorsalis on the humerus. Calonectris and Ardenna have a of the dorsal condyle and the open intercondylar furrow – larger ulna to carpometacarpus length ratio than Puffinus. that set it aside from all other species of shearwater. Ardenna Walker et al. (1990) found a similar relationship, where davealleni differs from P. inceptor in this way also. diving shearwater species had relatively longer outer wing Puffinus nestori Alcover, 1989 (Late Pleistocene, Pityusae bones. Calonectris species differ from species of Ardenna and Islands, Spain), P. holeae Walker, Wragg & Harrison, 1990 Puffinus in having a straighter femur shaft. Additionally, (Late Pleistocene, Canary Islands, Spain) and P. olsoni Calonectris and Ardenna are all larger than taxa of Puffinus. McMinn, Jaune & Alcover, 1990 (Holocene, Canary Islands) The extant C. diomedea and extinct C. wingatei are similar are all part of Puffinus (Heidrich et al. 1998), as defined in in size to A. davealleni, whereas the extant C. edwardsii and this paper, so are not closely related to Ardenna davealleni. extinct C. kurodai are smaller than A. davealleni (see Olson Puffinus spelaeus Holdaway & Worthy, 1994, from the 2008, 2009). Holocene of New Zealand, differs substantially from A. In conclusion, Ardenna davealleni shares most features with the species of Ardenna evolved for gliding (A. pacifica, davealleni as it is a close relative of P. gavia (see Tennyson & A. bulleri, A. carneipes, A. creatopus, A. gravis). Ardenna Shepherd 2017). davealleni differs from living shearwater taxa that are the Puffinus conradi Marsh, 1870 (Middle Miocene, most accomplished diving species, i.e. the clade of small taxa , USA) is in the subgenus Ardenna as a close relative (Puffinus) and A. grisea and A. tenuirostris (e.g. see Appendix of A. gravis (Wetmore 1926; Kuroda 1954: 114; Olson 1985c, 4), in having a rounded humeral shaft, a deeper fossa 2009), therefore we refer it to the genus Ardenna following the brachialis, a relatively straighter femur shaft, and a relatively taxonomy used in this paper. Ardenna conradi could, therefore, shorter cnemial crest on the tibiotarsus. be a close relative of A. davealleni. Unfortunately, A. conradi 8 Tuhinga, Number 29 (2018)

is known by only the ‘distal two-fifths of a left humerus’ Chandler (1990) to the ‘subgenus Thyellodroma’ (i.e. sister (Wetmore 1926). Nevertheless, A. conradi differs from A. taxon to Ardenna pacifica and A. bulleri). Therefore, we assign davealleni in having only a ‘slight’ processus supracondylaris P. gilmorei to Ardenna following the taxonomy used in this dorsalis (Wetmore 1926), whereas A. davealleni has a paper. Chandler (1990) noted that A. gilmorei differed from moderately sized process (see Appendix 1). In fact, the broader A. pacifica, A. bulleri and A. carneipes in having a more broadly ectepicondylar region of A. conradi is less developed than in undercut surface below the humeral head, but the shape of A. gravis (see Wetmore 1926), yet this region of A. davealleni this surface in A. davealleni is similar to that in A. pacifica and closely resembles that of A. gravis. The dorsoventral width of A. carneipes. Ardenna gilmorei is considerably smaller than the distal end of the A. conradi holotype measures 13.4 mm A. davealleni, having a humerus 96.7 mm long, a humerus (Wetmore 1926), so it is narrower than that of A. davealleni shaft 3.8 mm deep × 5.6 mm wide, a humeral distal at 14.2 mm, yet the dorsoventral diameter of the shaft of breadth of 8.1 mm and a humeral width at the processus the A. conradi holotype at the ‘upper end’ measures 7.1 mm supracondylaris dorsalis of 10.1 mm (Chandler 1990), (Wetmore 1926), which is similar to the mid-shaft width of compared with 121.8 mm, 5.6 mm, 7.0 mm, 14.2 mm and A. davealleni, at 7.0 mm. This indicates a proportional shape 13.4 mm, respectively, for A. davealleni. The femur of A. difference between the species. The greater age of A. conradi gilmorei is similarly small: length 35.1 mm, proximal breadth and its location on the Atlantic Ocean also argue against it 7.5 mm, shaft depth 3.3 mm, shaft width 3.3 mm and distal and A. davealleni being conspecific. breadth 7.2 mm (Chandler 1990). This compares with A. Puffinus aquitanicus (Milne-Edwards, 1874) (Middle davealleni measurements of 44.1 mm, 10.2 mm, c. 4.5 mm, Miocene, France) is based on a damaged humerus (Milne- c. 4.4 mm and 10.7 mm, respectively. Therefore, A. gilmorei Edwards 1874) that was initially considered similar to is a similar size to A. bulleri (see Appendix 3) and we Calonectris borealis or Fulmarus (according to Kuroda 1954: conclude that it and A. davealleni are not conspecific. 114; see Tennyson 2010 for an explanation of the current Puffinus pacificoides Olson, 1975 (Pleistocene, St Helena) names for these taxa), but it has a more elongated fossa ‘belongs to the pacificus-bulleri species group’ (Olson brachialis than that found in Calonectris and a stouter shaft 1975). Puffinus pacificoides should, therefore, be attributed (Olson 2009). Puffinus aquitanicus is somewhat smaller than to Ardenna using the taxonomy followed in this paper. Ardenna davealleni, having an estimated total humeral length Measurements for A. pacificoides femora are: length 32.8– of c. 112 mm, shaft depth of 4.5 mm, shaft width of 5.5 36.3 mm, proximal breadth 8.0–8.7 mm, shaft width mm and distal breadth of c. 12 mm (estimated here) (Milne- 3.5 mm and distal breadth 7.8 mm (Olson 1975). This Edwards 1874). This compares with 121.4 mm, 5.6 mm, 7.0 compares with measurements for A. davealleni of 44.1 mm and 14.2 mm, respectively, for A. davealleni. The true mm, 10.2 mm, c. 4.4 mm and 10.7 mm, respectively. affinities of P. aquitanicus remain unresolved, but its size, age Therefore, A. pacificoides was a substantially smaller bird and European location argue against it being conspecific with than A. davealleni (see Appendix 3). Olson noted that the A. davealleni. femur of A. pacificoides differs from that of A. pacifica and A second shearwater species named from the same location A. bulleri by having a much more pronounced trochanteric in the same article, Puffinus antiquus (Milne-Edwards, 1874) ridge, which is longer and bulges out further laterally. Also, (Middle Miocene, France), is based on the proximal end of the head of the femur is distinctly flatter and less bulbous in a humerus (Milne-Edwards 1874; Brodkorb 1963b). Milne- A. pacificoides (see Olson 1975). In A. davealleni, the femora Edwards’s description (1874) does not clearly distinguish this are partly obscured in the surrounding matrix, but the species from many other shearwaters, but he considered the femoral shape appears to be very similar to that in A. pacifica humerus to be only slightly stouter than that of a Cape petrel and A. bulleri. Similarly, Olson notes a range of features of (Daption capense), yet the humerus of Ardenna davealleni is the tarsometatarsus of A. pacificoides that differ from those considerably larger than that of a Cape petrel. The size, found in A. pacifica and A. bulleri. Ardenna davealleni differs European location and older age of Puffinus antiquus argue from A. pacificoides in these tarsometatarsus features, being against it being closely related to A. davealleni. more similar to all living Ardenna species. The humeral According to Olson (2009), Puffinus gilmoreiChandler, condyles of A. pacificoides were considered more delicate 1990 (Late Pliocene, California) was correctly assigned by than those of A. bulleri and more similar to those of A. A new species of Pliocene shearwater (Aves: Procellariidae) from New Zealand 9 pacifica (see Olson 1975). In size, the humeral condyles of Discussion A. davealleni resemble those of A. bulleri more closely than those of A. pacifica. The humeral shaft of A. davealleni is The discovery of a new species of Ardenna shearwater from stouter than that in A. bulleri, and in this way more closely the Pliocene adds important new knowledge about the resembles the humerus of A. pacifica and A. pacificoides (see history of this group. Although a few Pliocene shearwaters Olson 1975). Since the initial description, fossils of a species have been described before (Puffinus tedfordi, P. felthami, similar to A. pacificoides have been reported from an Early P. kanakoffi and A. gilmorei), all were found in the eastern Pliocene deposit in (Olson & Rasmussen Pacific (Howard 1949, 1971; Chandler 1990). While New 2001). Finally, comparisons with Puffinus raemdonckii (van Zealand today is the worldwide centre of diversity for species Beneden, 1871) of the Early Oligocene of Belgium of shearwater (Dickinson & Remsen 2013), A. davealleni are hampered by the fact that the type material has provides the only clear record of a shearwater genus for the apparently been lost, and so the true relationships region prior to the Late Pleistocene (Tennyson 2010) and between the species and Ardenna davealleni remain unclear demonstrates a long history for the genus in the western (Brodkorb 1962; Olson 1985c; Mayr & Smith 2012). Pacific. Many southern hemisphere shearwater species However, the lectotype of the species is a c. 10 cm long migrate to the northern hemisphere in their non-breeding humerus (missing the proximal end) (van Beneden 1871; season (e.g. Marchant & Higgins 1990), so it is possible that Brodkorb 1962), which is smaller than the humerus of northern hemisphere fossils represent such migrants (e.g. the A. davealleni. In addition, the much greater age of P. southern nesting A. grisea has been reported from an English raemdonckii argues against it being closely related to A. archaeological deposit; Harrison 1980). davealleni. Ardenna davealleni was a large species, structurally most Various other Miocene and Pliocene ‘Puffinus’ remains similar to the much smaller A. bulleri, which breeds only in have been noted in the literature but not assigned to a taxon New Zealand (Marchant & Higgins 1990). It is possible that (e.g. Howard 1968; Olson 1985a,b; Howard & Barnes 1987; A. davealleni is an ancestral form of this extant species. Based Chandler 1990; Olson & Rasmussen 2001; Hoffmeister on the behaviour of A. bulleri and its close relatives, A. et al. 2014). The only described Pliocene New Zealand davealleni was presumably more of a gliding species than a shearwater (a skull) is from a slightly older deposit than the diving bird. Taranaki fossil and is from a smaller species, closer in size Using a molecular clock, Penhallurick and Wink (2004) to Ardenna pacifica, but its true generic affinities remain considered that Calonectris shearwaters separated from other undetermined (Henderson & Gill 2010). shearwater clades c. 13.8 million years ago, and that Puffinus Note that the Late Miocene Pterodromoides minoricensis and Ardenna diverged c. 10.4 million years ago. Gómez- is considerably smaller than Ardenna davealleni and has Díaz et al. (2006) considered the Calonectris and ‘Puffinus’ cranial, coracoid and humeral features most similar to those of fulmarine petrels. Therefore, it is not closely related to shearwaters to have separated 9 million years ago, but shearwaters (Seguí et al. 2001). Olson (2009) noted that this may simply reflect the oldest In summary, many fossil shearwater taxa align with the divergence between the crown clades because the oldest genus Puffinus rather than Ardenna, owing to their relatively ‘Puffinus’ fossil is Oligocene and the oldest Calonectris fossil flattened humerus or their particularly laterally flattened is Middle Miocene. Our finding of a Pliocene Ardenna tarsometatarsus, so are not closely related to A. davealleni. shearwater (combined with the Californian Pliocene record Other fossil shearwater species differ from A. davealleni of A. gilmorei), well differentiated from species of Calonectris primarily through a combination of differences in humeral and Puffinus, adds weight to the theory of an early divergence shape and size, and the fact that they occur in different eras. of the three shearwater genera. 10 Tuhinga, Number 29 (2018)

Acknowledgements References We thank Dave Allen and Xavier and Alastair Johnson for Alcover, J.A. (1989). Les aus fòssils de la Cova de Ca Na Reia. providing their specimens to the Museum of New Zealand Endins 14–15: 95–100. Te Papa Tongarewa, Wellington, New Zealand (Te Papa) for Austin, J.J. (1996). Molecular phylogenetics of Puffinus research; Bruce Marshall (Te Papa) for providing information shearwaters: preliminary evidence from mitochondrial about molluscs; Ricardo Palma (Te Papa) for nomenclatural cytochrome b gene sequences. Molecular Phylogenetics and advice on the manuscript; Christine Kiddey and Martin Evolution 6: 77–88. Austin, J.J., Bretagnolle, V. and Pasquet, E. (2004). A global Lewis (Te Papa) for sourcing some references; and Jean- molecular phylogeny of the small Puffinus shearwaters Claude Stahl (Te Papa) for the photos and translation of and implications for systematics of the little-Audubon’s a foreign paper. We also thank Gerald Mayr (Senckenberg shearwater complex. Auk 121: 847–864. Research Institute and Natural History Museum Frankfurt, Baumel, J.J. and Witmer, L.M. (1993). Pp. 45–132. Frankfurt am Main, Germany), Hanneke Meijer (University Osteologia. In: Handbook of avian anatomy: nomina of Bergen, Norway), Nic Rawlence (Otago University, anatomica avium. 2nd edn. Baumel, J.J., King, A.S., , New Zealand), Brian Gill (, New Breazile, J.E., Evans, H.E. and Vanden Berge, J.C. Zealand) and Trevor Worthy (Flinders University, Adelaide, (eds). Publications of the Nuttall Ornithological Club ) for many helpful suggestions on ways to improve 23. Cambridge, MA: Nuttall Ornithological Club. xxiv the manuscript. + 779 pp. Brodkorb, P. (1962). The systematic position of two Oligocene birds from Belgium. Auk 79: 706–707. Nomenclatural act Brodkorb, P. (1963a). Miocene birds from the Hawthorne Formation. Quarterly Journal of the Florida Academy of This published work and the nomenclatural act it Sciences 2: 159–167. contains have been registered in ZooBank, the online Brodkorb, P. (1963b). Catalogue of fossil birds – part 1 registration system for the International Code of Zoological (Archaeopterygiformes through Ardeiformes). Bulletin of Nomenclature. The ZooBank Life Science Identifier the Florida State Museum 7: 179–293. for the taxon in this publication is http://zoobank.org/ Campbell, K.E. and Marcus, L. (1992). The relationship of urn:lsid:zoobank.org: act: 3F25CA99-540F-41E5-AD3B- hindlimb bone dimensions to body weight in birds. DA533EC2E6C2. Pp. 395–412. In: Papers in paleontology – honoring Pierce Brodkorb: incorporating the Proceedings of the II International Symposium of the Society of Avian Paleontology and Evolution held at the Natural History Museum of Los Angeles County, 28–30 September 1988. Campbell, K.E. (ed.). Science Series 36. Los Angeles, CA: Natural History Museum of Los Angeles County. 491 pp. Chandler, R.M. (1990). Recent advances in the study of Neogene fossil birds. Part II. Fossil birds of the San Diego Formation, Late Pliocene, Blancan, San Diego County, California. Ornithological Monographs 44: 73–161. Dickinson, E.C. and Remsen, J.V. (eds) (2013). The Howard and Moore complete checklist of the birds of the world. Vol. 1. 4th edn. Eastbourne: Aves Press. l + 461 pp. Fleming, C.A. (1979). The geological and its life. Auckland: Auckland University Press. 141 pp. Fordyce, R.E. (1991). A new look at the fossil vertebrate record of New Zealand. Pp. 1191–1316. In: Vertebrate palaeontology of Australasia. Vickers-Rich, P., Monaghan, J.M., Baird, R.F. and Rich, T.H. (eds). Melbourne: Pioneer Design Studio. 1437 pp. A new species of Pliocene shearwater (Aves: Procellariidae) from New Zealand 11

Gómez-Díaz, E., González-Solís, J., Peinado, M.A. and Howard, H. and Barnes, L.G. (1987). Middle Miocene marine Page, R.D.M. (2006). Phylogeography of the Calonectris birds from the foothills of the Santa Ana Mountains, Orange shearwaters using molecular and morphometric data. County, California. Natural History Museum of Los Molecular Phylogenetics and Evolution 41: 322–332. Angeles County Contributions in Science 383. Los Harrison, C.J.O. (1980). A re-examination of British Angeles, CA: Natural History Museum of Los Angeles Devensian and earlier Holocene bird bones in the British County. 9 pp. Museum (Natural History). Journal of Archaeological Howard, H. and Warter, S.L. (1969). A new species of bony- Science 7: 53–68. toothed bird (Family Pseudodonthornithidae [sic]) from Harrison, C.J.O. and Walker, C.A. (1976). A review of the the Tertiary of New Zealand. Records of Canterbury bony-toothed birds (Odontopterygiformes): with descriptions Museum 8: 345–357. of some new species. Tertiary Research Special Paper 2. Ksepka, D.T. and Cracraft, J. (2008). An avian tarsometatarsus London: Tertiary Research Group. 72 pp. from near the K–T boundary of New Zealand. Journal of Heidrich, P., Amengual, J. and Wink, M. (1998). Vertebrate Paleontology 28: 1224–1227. Phylogenetic relationships in Mediterranean and Ksepka, D.T., Fordyce, R.E., Ando, T. and Jones, C.M. (2012). North Atlantic shearwaters (Aves: Procellariidae) based New fossil penguins (Aves, Sphenisciformes) from the on nucleotide sequences of mt DNA. Biochemical Oligocene of New Zealand reveal the skeletal plan of stem Systematics and Ecology 26: 145–170. penguins. Journal of Vertebrate Paleontology 32: 235–254. Henderson, N. and Gill, B.J. (2010). A mid-Pliocene shearwater Kuroda, N. (1953). On the skeletons of Puffinus nativitatis skull (Aves: Procellariidae: Puffinus) from the Taihape and Pagodroma nivea. To r i 13: 50–68. Mudstone, central North Island, New Zealand. New Kuroda, N. (1954). On the classification and phylogeny of the Zealand Journal of Geology and Geophysics 53: 327–332. order tubinares, particularly the shearwaters (Puffinus). Hoffmeister, M.C., Carrillo, J.D. and Nielsen, S.N. (2014). Tokyo: N. Kuroda. 179 pp. The evolution of seabirds in the Humboldt Current: Kuroda, N. (1955). On the osteology of the gadfly-petrels, new clues from the Pliocene of central Chile. PloS ONE Pterodroma. Annotationes Zoologicae Japonenses 28: 171–177. 9(3): e90043. doi:10.1371/journal.pone.0090043. Leach, W.E. (1820). Eleventh room, upper floor. Pp. 65–70. Holdaway, R.N. and Worthy, T.H. (1994). A new fossil In: Synopsis of the contents of the British Museum. 17th edn. species of shearwater Puffinus from the Late Quaternary London: British Museum. 162 pp. of the , New Zealand, and notes on the Livezey, B.C. and Zusi, R.L. (2006). Higher-order phylogeny of biogeography and evolution of the Puffinus gavia modern birds (Theropoda, Aves: Neornithes) based on superspecies. Emu 94: 201–215. comparative anatomy. I. Methods and characters. Bulletin of Howard, H. (1949). New avian records for the Pliocene of the Carnegie Museum of Natural History 37. Pittsburgh, California. Pp. 179–199. In: Some Tertiary mammals and PA: Carnegie Museum of Natural History. 544 pp. birds from North America. Carnegie Institution of McKee, J.W.A. (1985). A pseudodontorn (: Washington Publication 584. Washington, DC: ) from the middle Pliocene of Hawera, Carnegie Institution of Washington. 244 pp. Taranaki, New Zealand. New Zealand Journal of Zoology Howard, H. (1968). Tertiary birds from Laguna Hills, 12: 181–184. Orange County, California. Natural History Museum McKee, J.W.A. (1988). The occurrence of the Pliocene of Los Angeles County Contributions in Science 142. penguin Tereingaornis moisleyi (Sphenisciformes: Los Angeles, CA: Natural History Museum of Los Spheniscidae) at Hawera, Taranaki, New Zealand. New Angeles County. 21 pp. Zealand Journal of Zoology 14: 557–561. Howard, H. (1971). Pliocene avian remains from Baja California. McKee, J.W.A. (1994). Geology and vertebrate palaeontology Natural History Museum of Los Angeles County of the Tangahoe Formation, South Taranaki coast, New Contributions in Science 217. Los Angeles, CA: Natural Zealand. Geological Society of New Zealand Miscellaneous History Museum of Los Angeles County. 17 pp. Publication 80B: 63–91. Howard, H. (1978). Late Miocene marine birds from Orange McMinn, M., Jaume, D. and Alcover, J.A. (1990). Puffinus County, California. Natural History Museum of Los olsoni n. sp.: nova espècie de baldritja recentment Angeles County Contributions in Science 290. Los extingida provinent de depòsits espeleològics de Angeles, CA: Natural History Museum of Los Angeles Fuerteventura i Lanzarote (Illes Canàries, Atlàntic County. 26 pp. Oriental). Endins 16: 63–71. 12 Tuhinga, Number 29 (2018)

Marchant, S. and Higgins, P.J. (coordinators) (1990). Handbook Olson, S.L. (1985b). An Early Pliocene marine avifauna of Australian, New Zealand and Antarctic birds. Vol. 1: from Duinefontein, Cape Province, South Africa. ratites to ducks. Melbourne: Oxford University Press. Annals of the South African Museum 95: 147–164. 1400 pp. Olson, S.L. (1985c). The fossil record of birds. Avian Marsh, O.C. (1870). Notice of some fossil birds, from the Biology 8: 79–238. Cretaceous and Tertiary formations of the United States. Olson, S.L. (2008). A new species of shearwater of the genus American Journal of Science and Arts, Series 2 49: 205–217. Calonectris (Aves: Procellariidae) from a Middle Mayr, G. and Scofield, R.P. (2014). First diagnosable non- Pleistocene deposit on Bermuda. Proceedings of the sphenisciform bird from the Early Paleocene of New Biological Society of Washington 121: 398–409. Zealand. Journal of the Royal Society of New Zealand Olson, S.L. (2009). A new diminutive species of shearwater of 44: 48–56. the genus Calonectris (Aves: Procellariidae) from the Mayr, G. and Scofield, R.P. (2015). New avian remains from Middle Miocene Calvert Formation of Chesapeake Bay. the Paleocene of New Zealand: the first early Cenozoic Proceedings of the Biological Society of Washington 122: Phaethontiformes (tropicbirds) from the southern 466–470. hemisphere. Journal of Vertebrate Paleontology 36: Olson, S.L. and Rasmussen, P.C. (2001). Miocene and e1031343. doi:10.1080/02724634.2015.1031343. Pliocene birds from the Lee Creek Mine, North Carolina. Mayr, G. and Smith, T. (2012). Phylogenetic affinities and Pp. 233–365. In: Geology and paleontology of the Lee Creek Mine, North Carolina, III. Ray, C.R. and Bohaska, taxonomy of the Oligocene Diomedeoididae, and the D.J. (eds). Smithsonian Contributions to Paleobiology 90. basal divergences amongst extant procellariiform birds. Washington, DC: Smithsonian Institution Press. 365 pp. Zoological Journal of the Linnean Society 166: 854–875. Onley, D. and Scofield, P. (2007).Albatrosses, petrels, and Mayr, G., Scofield, R.P., De Pietri, V.L. and Tennyson, shearwaters of the world. Princeton, NJ: Princeton A.J.D. (2017). A Paleocene penguin from New Zealand University Press. 240 pp. substantiates multiple origins of gigantism in fossil O’Reilly, B. (1818). Greenland, the adjacent seas, and the Sphenisciformes. Nature Communications 8: 1927 (2017). North-west Passage to the Pacific Ocean, illustrated in a doi:10.1038/s41467-017-01959-6. voyage to Davis’s Strait, during the summer of 1817. Miller, L. (1925). Avian remains from the Miocene of London: Baldwin, Cradock, and Joy. 293 pp. Lompoc, California. Carnegie Institution of Washington Penhallurick, J. and Wink, M. (2004). Analysis of the 349: 107–117. taxonomy and nomenclature of the Procellariiformes Miller, L. (1961). Birds from the Miocene of Sharktooth Hill, based on complete nucleotide sequences of the California. Condor 63: 399–402. mitochondrial cytochrome b gene. Emu 104: 125–147. Milne-Edwards, A. (1874). Observations sur les oiseaux Peters, S.E. and Loss, D.P. (2012). Storm and fair-weather fossiles des Faluns de Saucats. Bibliothèque de l’École des wave base: a relevant distinction? Geology 40: 511–514. Hautes Études, Section des Sciences naturelles 11(3): 1–12. Reichenbach, H.G.L. (1849–53). Avium systema naturale: Mourer-Chauviré, C. and Geraads, D. (2010). The Upper das natürliche System der Vögel mit hundert Tafeln Pliocene avifauna of Ahl al Oughlam, Morocco. grosstentheils Original-Abbildungen der bis jetzt entdecken Systematics and biogeography. Records of the Australian fast zwölfhundert typischen Formen. Vorläufer einer Museum 62: 157–184. Iconographie der Arten der Vögel aller Welttheile, Nunn, G.B. and Stanley, S.E. (1998). Body size effects and welche, nachdem bereits fast dreitausend Abbildungen rates of cytochrome b evolution in tube-nosed seabirds. erscheinen sind, ununterbrochen fortgesetzt wird. Dresden Molecular Biology and Evolution 15: 1360–1371. and Leipzig: Expedition der vollstandigsten Olson, S.L. (1975). Paleornithology of St Helena Island, South Naturgeschichte. 100 pp. + 100 pls. Atlantic Ocean. Smithsonian Contributions to Scarlett, R.J. (1972). Bone of a presumed odontopterygian Paleobiology 23. Washington, DC: Smithsonian Institution bird from the Miocene of New Zealand. New Zealand Press. iv + 49 pp. Journal of Geology and Geophysics 15: 269–274. Olson, S.L. (1985a). Early Pliocene Procellariiformes (Aves) Scofield, R.P., Hiller, N. and Mannering, A.A. (2006). from Langebaanweg, South-Western Cape Province, A fossil diving petrel (Aves: Pelecanoididae) from the South Africa. Annals of the South African Museum 95: Mid-Miocene of North Canterbury, New Zealand. 123–145. Records of the Canterbury Museum 20: 65–71. A new species of Pliocene shearwater (Aves: Procellariidae) from New Zealand 13

Seguí, B., Quintana, J., Fornós, J.J. and Alcover, J.A. (2001). Medway, D.G., Palma, R.L., Scofield, R.P., Tennyson, A new fulmarine petrel (Aves: Procellariiformes) from the A.J.D. and Worthy, T.H. (eds). Wellington: Ornithological Upper Miocene of the western Mediterranean. Society of New Zealand and Te Papa Press. 500 pp. Palaeontology 44: 933–948. Worthy, T.H., Tennyson, A.J.D., Jones, C., McNamara, J.A. Simpson, G.G. (1975). Fossil penguins. Pp. 19–41. In: The and Douglas, B.J. (2007). Miocene waterfowl and other biology of penguins. Stonehouse, B. (ed.). London: birds from Central Otago, New Zealand. Journal of Macmillan Press. 555 pp. Systematic Palaeontology 5(1): 1–39. Tennyson, A.J.D. (2010). Procellariiformes. Pp. 64–135. In: Checklist of the birds of New Zealand, Norfolk and Macquarie Islands, and the Ross Dependency, Antarctica. 4th edn. Gill, B.J., Bell, B.D., Chambers, G.K., Medway, D.G., Palma, R.L., Scofield, R.P., Tennyson, A.J.D. and Worthy, T.H. (eds). Wellington: Ornithological Society of New Zealand and Te Papa Press. 500 pp. Tennyson, A.J.D. and Shepherd, L.D. (2017). DNA reveals the relationships of the extinct Scarlett’s shearwater Puffinus spelaeus (Procellariiformes: Procellariidae Journal of Ornithology 158: 379–384. van Beneden, P.J. (1871). Les oiseaux de l’argile rupelienne. Bulletins de l’Académie royale des sciences, des lettres et des beaux-arts de Belgique Ser. 2 32(11): 256–261. Walker, C.A., Wragg, G.M. and Harrison, C.J.O. (1990). A new shearwater from the Pleistocene of the Canary Islands and its bearing on the evolution of certain Puffinus shearwaters. Historical Biology 3: 203–224. Warham, J. (1990). The petrels: their ecology and breeding systems. London: Academic Press. viii + 440 pp. Warham, J. (1996). The behaviour, population biology and physiology of the petrels. London: Academic Press. vii + 613 pp. Warheit, K.I. (2002). The seabird fossil record and the role of paleontology in understanding seabird community structure. Pp. 17–55. In: Biology of marine birds. Schreiber, E.A. and Burger, J. (eds). Boca Raton, FL: CRC Press. 740 pp. Wetmore, A. (1926). Observations on fossil birds described from the Miocene of Maryland. Auk 43: 462–468. Wetmore, A. (1930). Fossil bird remains from the Temblor Formation near Bakersfield, California.Proceedings of the California Academy of Sciences 19: 85–93. Worthy, T.H. and Holdaway, R.N. (2002). The lost world of the : prehistoric life of New Zealand. Christchurch: Canterbury University Press. xxxii + 718 pp. Worthy, T.H. and Tennyson, A.J.D. (2010). Appendix 3: fossil birds of New Zealand. Pp. 323–333. In: Checklist of the birds of New Zealand, Norfolk and Macquarie Islands, and the Ross Dependency, Antarctica. 4th edn. Gill, B.J., Bell, B.D., Chambers, G.K., 14 Tuhinga, Number 29 (2018)

Appendix 1: Data matrix of scored morphological characters of Procellariiformes as described in Appendix 2

Character Species 1 2 3 4 5 6 7 8 9 10 11 12 13 14

Ardenna davealleni n. sp. S.45183 1 1 1 1 0 2 1 1 2 1 2 2 3 4 and S.46316 Phoebastria, e.g. P. immutabilis 0 0 0 2 0 2 2 0 0 1 5 5 5 5 Macronectes, e.g. M. halli 1 1 0 2 0 2 0 0 0 2 4 4 3 5 Fulmarus, e.g. F. glacialoides 2 1 0 2 0 2 0 0 0 1 3 2 2 4 Thalassoica antarctica 2 1 0 2 0 2 0 0 1 1 3 2 1 3 Daption capense 2 1 1 2 0 2 0 1 1 1 3 2 2 3 Pagodroma nivea 2 1 1 1 0 2 0 0 1 2 2 2 1 3 Lugensa brevirostris 2 2 0 3 0 2 0 1 1 0 3 3 3 3 Pterodroma, e.g. Pt. macroptera 2 0 0 2 0 3 0 1 1 0 3 3 3 2–4 Halobaena caerulea 2 0 0 2 0 2 0 1 1 1 2 2 2 2 Pachyptila, e.g. P. vittata 2 0 0 2 0 2 0 1 1 1 2 3 2 1–2 Bulweria, e.g. B. bulwerii 2 1 1 2 0 2 0 1 0 0 3 3 4 2 Procellaria, e.g. P. parkinsoni 1 1 0 2 0 3 0 1 1 1 3 3 3 4 Pseudobulweria, e.g. P. rostrata 1 1 1 2 0 3 0 1 ?0 1 3 3 – 2–3 Calonectris borealis 1 1 1 2 0 3 1 1 1 1 3 3 3 4 Calonectris leucomelas 1 1 1 2 0 3 1 1 1 1 3 3 3 4 Ardenna pacifica 1 1 1 1 0 2 1 1 2 0 3 4 4 4 Ardenna bulleri 2 1 1 1 0 2 1 1 2 1 2 3 3 4 Ardenna carneipes 1 1 1 2 0 3 1 1 2 1 2 3 3 4 Ardenna creatopus 1 1 1 2 0 3 1 1 2 1 2 3 3 4 Ardenna gravis 1 1 1 2 0 3 1 1 2 1 2 3 3 4 Ardenna grisea 1 1 1 1 1 2 2 1 3 2 2 3 3 4 Ardenna tenuirostris 1 1 1 1 1 2 2 1 3 2 2 3 3 4 Puffinus nativitatis 1 1 1 1 1 1 2 1 3 1 1 3 3 3 Puffinus puffinus* 1 1 1 1 1 1 2 1 4 2 1 2 2 3 Puffinus assimilis* 1 1 1 1 1 1 2 1 4 2 1 2 2 1–2 Northern storm petrel*, 2 2 1 0 0 0 0 0 0 1 1 1 2 1 e.g. Oceanodroma leucorhoa Southern storm petrel*, 2 2 1 2 0 0 0 0 1 1 0 0 0 1 e.g. Nesofregetta fuliginosa Pelecanoides, e.g. P. urinatrix 2 2 1 0 1 0 2 1 2 3 0 1 0 1

*Manx-type shearwater group, little/Audubon’s shearwater complex and northern and southern storm petrel groups follow Onley & Scofield (2007). A new species of Pliocene shearwater (Aves: Procellariidae) from New Zealand 15

Appendix 2: List of morphological characters of Procellariiformes used in this analysis

The list comprises a of newly derived characters and those from Kuroda (1953, 1954, 1955), Chandler (1990), Olson & Rasmussen (2001) and Olson (2008, 2009). Ratio groups were chosen to demonstrate proportional differences between taxa. The positions of most characters are shown in Fig. 1. Character scores for Ardenna davealleni n. sp. NMNZ S.45183 and/or NMNZ S.46316 are shown in bold.

1 Humerus, proximal end, caput humeri undercut and overhangs fossa pneumotricipitalis: not at all (0); slightly (1); considerably (2). The characters of the proximal end of the humerus described by Seguí et al. (2001) are directly linked with characters 1 and 2 in this paper.

2 Humerus, proximal end, capital shaft ridge: does not extend to caput humeri (0); extends to caput humeri (1); extends to tuberculum dorsale (2).

3 Humerus, proximal end, bicipital area, tubercle on the cranial surface of the ventral margin: raised (0); not present, or only slightly raised (1).

4 Humerus, proximal end, crista deltopectoralis: very low rounded angle (0); rounded angle (1); sharper angle/point (2); sharper point with strongly concave outline distal of point (3).

5 Humerus shaft: rounded, with ratio of central shaft width to depth <1.5 (0); dorsoventrally flattened, with ratio of central shaft width to depth >1.5 (1). Character 5 is also associated with the ventral epicondylar area being expanded. Note that this character does not separate all other shearwaters from Calonectris (contra Mourer-Chauviré & Geraads 2010) and a flattened humerus shaft is not an apomorphy of shearwaters (contra Seguí et al. 2001), although this feature is found in all Puffinus as defined in this paper.

6 Humerus, distal end, processus supracondylaris dorsalis (= ectepicondylar process): small (0); moderately sized, angled distinctly proximally (1); moderately sized but projects straight (2); large straight process (3).

7 Humerus, distal end, fossa m. brachialis: deep (0); medium depth (1); shallow (2).

8 Proximal end of the ulna, caudal surface of the proximal shaft forms a prominent and continuous ridge to the tip of the proximal end: absent (0); present (1).

9 Femur shaft medial view: fairly straight (0); slightly curved (1); fairly curved (2); very curved (3); extremely curved (4). Curvature is arching dorsally.

10 Ratio of humerus length to ulna length: <0.98 (0); 0.98–1.04 (1); 1.05–1.19 (2); >1.19 (3). Kuroda (1955) considered gliding species to have a longer ulna than humerus.

11 Ratio of ulna length to carpometacarpus length: <1.58 (0); 1.59–1.79 (1); 1.80–1.94 (2); 1.95–2.24 (3); 2.25–2.45 (4); >2.45 (5).

12 Ratio of ulna length to manus phalanx proximal digitus major length: <2.69 (0); 2.69–3.25 (1); 3.26–3.65 (2); 3.66–4.19 (3); 4.20–4.73 (4); >4.73 (5).

13 Ratio of ulna length to femur length: <1.61 (0); 1.61–2.09 (1); 2.10–2.44 (2); 2.45–3.07 (3); 3.08–3.39 (4); >3.39 (5).

14 Overall size (length): very small, <26 cm long (1); small, 26–32 cm long (2); medium, 33–41 cm long (3); large, 42–55 cm long (4); very large, >74 cm long (5) (scores based on Onley & Scofield 2007 plates).

16 Tuhinga, Number 29 (2018)

Appendix 3: Measurements of Procellariidae bones and ratios of these bones

Measurements are total lengths in millimetres. Abbreviations: Hum = humerus, Uln = ulna, Rad = radius, Cmc = carpometacarpus, Man = manus phalanx proximal digitus major, Fem = femur. Data for Ardenna species based on means presented in Appendix 5.

Taxa Hum Uln Rad Cmc Man Fem Hum/Uln Uln/Cmc Uln/Man Uln/Fem

Ardenna davealleni n. sp. – – NMNZ S.45183 113.4 111.4 62.9 31.8 44.1 1.81 3.57 2.57 Ardenna davealleni n. sp 121.8 117.7 112.8 – – – 1.03 – – – NMNZ S.46316 Phoebastria immutabilis OR.22147 254 254 251 99.4 49.3 71.9 1.00 2.56 5.15 3.53 Macronectes halli OR.28597 257 242 235 101.8 56.3 91.3 1.06 2.34 4.30 2.65 Fulmarus glacialoides OR.29079 115.0 111.7 109.0 55.5 31.4 50.5 1.03 2.01 3.56 2.21 Thalassoica antarctica OR.27407 85.6 85.9 83.5 43.9 25.2 43.3 1.00 1.96 3.41 1.98 Daption capense OR.29529 83.3 81.9 79.8 40.7 22.8 37.1 1.02 2.01 3.59 2.21 Pagodroma nivea OR.11947 78.5 73.4 71.5 39.3 21.7 41.8 1.07 1.87 3.38 1.76 Lugensa brevirostris OR.29066 77.5 82.1 80.5 39.8 21.0 32.4 0.94 2.06 3.91 2.53 Pterodroma macroptera OR.27820 109.5 113.1 110.5 54.1 28.4 39.8 0.97 2.09 3.98 2.84 Halobaena caerulea OR.17591 61.4 59.0 57.5 31.8 17.3 25.8 1.04 1.86 3.41 2.29 Pachyptila vittata OR.27311 61.0 61.4 59.5 32.2 16.2 26.2 0.99 1.91 3.79 2.34 Bulweria bulwerii OR.22145 60.4 62.4 61.1 30.0 16.2 19.2 0.97 2.08 3.85 3.25 Procellaria parkinsoni OR.25953 129.8 130.6 127.9 63.6 33.2 43.9 0.99 2.05 3.93 2.97 Pseudobulweria rostrata OR.23900 109.0 108.5 106.6 51.7 27.1 – 1.00 2.10 4.00 – Calonectris borealis OR.27733 128.2 128.8 126.1 62.9 33.4 43.9 1.00 2.05 3.86 2.93 Calonectris leucomelas OR.29195 112.6 114.1 110.8 54.9 29.3 39.1 0.99 2.08 3.89 2.92 Ardenna pacifica 102.5 106.8 103.9 52.3 25.5 34.1 0.96 2.04 4.20 3.13 Ardenna bulleri 96.6 97.7 94.6 50.4 25.3 35.4 0.99 1.94 3.86 2.76 Ardenna carneipes 112.6 113.1 107.9 59.2 29.5 40.7 1.00 1.91 3.83 2.79 Ardenna creatopus 114.3 112.3 108.7 59.7 29.9 41.8 1.01 1.88 3.82 2.69 Ardenna gravis 119.8 117.6 113.9 61.7 30.2 39.9 1.03 1.91 3.89 2.84 Ardenna grisea 107.1 100.3 96.5 55.7 27.4 39.8 1.07 1.80 3.66 2.51 Ardenna tenuirostris 97.9 92.2 90.0 49.7 24.3 35.7 1.06 1.86 3.80 2.59 Puffinus nativitatisOR .24682 78.2 78.3 75.3 45.1 21.0 28.4 1.00 1.74 3.73 2.76 Puffinus puffinusOR .26859 78.4 70.8 68.3 42.5 21.2 30.7 1.11 1.66 3.34 2.31 Puffinus assimilisOR .23972 62.8 56.4 54.3 33.1 16.4 24.9 1.11 1.70 3.44 2.27 Oceanodroma leucorhoa OR.19320 31.9 32.2 30.8 18.1 10.2 14.7 0.99 1.78 3.16 2.19 Nesofregetta fuliginosa OR.27482 31.2 30.9 29.7 20.5 12.9 21.3 1.01 1.51 2.40 1.45 Pelecanoides urinatrix OR.28939 43.6 34.7 32.1 25.0 11.7 23.8 1.26 1.39 2.97 1.46 A new species of Pliocene shearwater (Aves: Procellariidae) from New Zealand 17

Appendix 4: Measurements of legs bones of Ardenna davealleni n. sp. paratype NMNZ S.46316 compared with other Ardenna shearwater species

Measurements are maximum lengths in millimetres. Where known, sex is shown as M = male, F = female. The tibiotarsus (tbt) measurements are presented with and without the cnemial crest. The proportion of the cnemial crest to the full length of the tibiotarsus is presented as % crest length. Tmt = tarsometatarsus.

Taxa Tbt + crest Tbt – crest % crest length Tmt Taxa Tbt + crest Tbt – crest % crest length Tmt Ardenna davealleni n.sp. Ardenna creatopus NMNZ OR.26623 F 94.7 81.7 13.7 56.3 S.46316 108.0 90.8 15.9 60.7 OR.27754 91.7 78.7 14.2 54.0 Ardenna pacifica OR.27755 95.0 81.3 14.4 56.7 OR.16209 88.8 77.3 13.0 54.7 OR.27756 96.2 82.4 14.3 56.3 OR.21455 M 80.1 70.2 12.4 49.4 Mean 94.4 81.0 14.2 55.8 OR.23001 F 78.2 68.6 12.3 48.0 Ardenna gravis OR.28441 F 82.1 72.2 12.1 50.7 OR.22144 105.3 89.8 14.7 60.9 OR.29669 M 78.4 68.4 12.8 48.7 Ardenna grisea OR.29670 F 79.2 69.0 12.9 48.4 OR.13311 F 100.3 80.2 20.0 53.7 Mean 81.1 71.0 12.6 50.0 OR.13312 F 106.0 85.2 19.6 57.2 Ardenna bulleri OR.15619 M 105.2 85.7 18.5 56.8 OR.29075 81.8 71.1 13.1 50.1 OR.15620 M 103.8 85.7 17.4 58.5 OR.29479 F 82.4 71.3 13.5 49.9 OR.15922 M 106.1 86.1 18.9 57.9 OR.29496 M 81.6 70.5 13.6 51.7 OR.29656 M 102.4 84.0 18.0 56.6 OR.29625 M 80.8 69.5 14.0 48.6 Mean 104.0 84.5 18.7 56.8 OR.29723 M 82.9 71.5 13.8 49.7 Ardenna tenuirostris OR.29993 82.3 71.9 12.6 49.4 OR.14987 M 94.5 77.9 17.6 52.2 Mean 82.0 71.0 13.4 49.9 OR.15908 90.9 75.3 17.2 49.7 Ardenna carneipes OR.25134 89.4 74.0 17.2 47.5 OR.15924 88.0 75.6 14.1 53.4 OR.25557 95.9 79.5 17.1 52.0 OR.24666 93.4 80.3 14.0 56.6 OR.29306 94.1 77.8 17.3 51.3 OR.26223 95.4 82.9 13.1 57.7 OR.29471 F 97.9 81.1 17.2 53.5 OR.26447 94.5 80.6 14.7 58.3 Mean 93.8 77.6 17.3 51.0 OR.29119 89.6 76.7 14.4 54.8 OR.29207 M 90.5 77.0 14.9 53.9 Mean 91.9 78.9 14.2 55.8 18 Tuhinga, Number 29 (2018)

Appendix 5: Measurements of Ardenna shearwater bones and ratios of these bones

Measurements are total lengths in millimetres. Where known, sex is shown as M = male, F = female. Abbreviations: Hum = humerus, Uln = ulna, Rad = radius, Cmc = carpometacarpus, Man = manus phalanx proximal digitus major, Fem = femur.

Hum/ Taxa Hum Uln Rad Cmc Man Fem Uln Uln/Cmc Uln/Man Uln/Fem Ardenna pacifica OR.16209 110.9 116.7 113.4 57.0 29.2 36.9 0.95 2.05 4.00 3.16 OR.21455 M 103.1 105.5 103.0 52.3 26.1 34.2 0.98 2.02 4.04 3.08 OR.23001 F 101.0 105.5 103.1 50.7 24.3 32.7 0.96 2.08 4.34 3.23 OR.28441 F 104.3 109.3 106.1 53.4 24.7 35.0 0.95 2.05 4.43 3.12 OR.29669 M 99.2 102.7 99.8 50.2 24.2 32.5 0.97 2.05 4.24 3.16 OR.29670 F 96.5 101.1 98.2 50.4 24.4 33.1 0.95 2.01 4.14 3.05 Mean 102.5 106.8 103.9 52.3 25.5 34.1 0.96 2.04 4.20 3.13

Ardenna bulleri OR.29075 97.5 98.0 94.9 50.2 24.4 34.8 0.99 1.95 4.02 2.82 OR.29479 F 96.9 96.8 93.9 48.8 25.0 35.8 1.00 1.98 3.87 2.70 OR.29496 M 97.4 97.6 94.5 52.4 26.6 35.5 1.00 1.86 3.67 2.75 OR.29625 M 97.7 100.2 97.3 50.2 24.5 34.4 0.98 2.00 4.09 2.91 OR.29723 M 96.7 98.2 94.2 51.9 26.5 35.5 0.98 1.89 3.71 2.77 OR.29993 93.5 95.6 92.6 49.1 25.0 36.5 0.98 1.95 3.82 2.62 Mean 96.6 97.7 94.6 50.4 25.3 35.4 0.99 1.94 3.86 2.76 Ardenna carneipes OR.15924 106.3 107.2 103.4 56.2 27.9 39.6 0.99 1.91 3.84 2.71 OR.24666 114.7 115.0 111.8 60.8 30.5 41.2 1.00 1.89 3.77 2.79 OR.26223 116.1 117.0 103.7 60.6 29.9 41.5 0.99 1.93 3.91 2.82 OR.26447 113.8 114.2 110.7 59.9 30.3 41.7 1.00 1.91 3.77 2.74 OR.29119 112.8 113.5 109.8 59.1 29.0 39.4 0.99 1.92 3.91 2.88 OR.29207 M 111.6 111.6 108.2 58.6 29.4 – 1.00 1.90 3.80 – Mean 112.6 113.1 107.9 59.2 29.5 40.7 1.00 1.91 3.83 2.79 Ardenna creatopus OR.26623 F 114.6 114.2 110.6 59.2 29.4 41.5 1.00 1.93 3.88 2.75 OR.27754 110.0 108.3 105.2 57.1 28.9 40.5 1.02 1.90 3.74 2.67 OR.27755 114.5 113.7 109.9 61.7 30.3 43.4 1.01 1.84 3.75 2.62 OR.27756 114.8 112.8 109.2 60.8 30.9 41.7 1.02 1.86 3.90 2.71 OR.29584 118.2 – – – – – – – – – OR.29585 113.8 – – – – – – – – – Mean 114.3 112.3 108.7 59.7 29.9 41.8 1.01 1.88 3.82 2.69 A new species of Pliocene shearwater (Aves: Procellariidae) from New Zealand 19

Hum/ Taxa Hum Uln Rad Cmc Man Fem Uln Uln/Cmc Uln/Man Uln/Fem Ardenna gravis OR.22144 120.8 117.6 113.9 61.7 30.2 41.4 1.03 1.91 3.89 2.84 OR.30073 F 118.8 – – – – 38.3 – – – – Mean 119.8 – – – – 39.9 – – – – Ardenna grisea OR.13311 F 102.6 93.8 89.8 52.6 26.5 39.1 1.09 1.78 3.54 2.40 OR.13312 F 108.9 103.0 99.6 56.5 28.0 39.7 1.06 1.82 3.68 2.59 OR.15619 M 107.4 100.2 96.1 56.4 27.5 40.8 1.07 1.78 3.64 2.46 OR.15620 M 107.0 102.0 98.5 56.6 27.6 38.7 1.05 1.80 3.70 2.59 OR.15922 M 109.3 103.6 99.9 58.0 28.3 40.8 1.06 1.79 3.66 2.54 OR.29656 M 107.1 98.9 95.2 54.3 26.5 39.7 1.08 1.82 3.73 2.49 Mean 107.1 100.3 96.5 55.7 27.4 39.8 1.07 1.80 3.66 2.51 Ardenna tenuirostris OR.14987 M 99.0 93.7 90.5 50.8 24.9 37.0 1.06 1.84 3.76 2.53 OR.15908 94.2 87.9 83.7 47.5 22.5 33.7 1.07 1.85 3.91 2.61 OR.25134 91.2 86.8 84.3 46.6 22.7 33.4 1.05 1.86 3.82 2.60 OR.25557 102.6 95.7 92.4 51.0 25.3 37.0 1.07 1.88 3.78 2.61 OR.29306 98.4 92.7 89.6 49.7 24.1 37.0 1.06 1.87 3.85 2.51 OR.29471 F 101.8 96.3 93.3 52.3 26.0 36.3 1.06 1.84 3.70 2.65 Mean 97.9 92.2 90.0 49.7 24.3 35.7 1.06 1.86 3.80 2.59 Tuhinga 29: 20–34 Copyright © Museum of New Zealand Te Papa Tongarewa (2018)

New locality records for two species of protected weevils, Anagotus fairburni (Brookes, 1932) and Hadramphus stilbocarpae Kuschel, 1971 (Coleoptera: Curculionidae), from southern Fiordland, New Zealand Colin M. Miskelly,* Alan J.D. Tennyson** and Colin R. Bishop***

* Museum of New Zealand Te Papa Tongarewa, PO Box 467, Wellington 6140, New Zealand ([email protected]) ** Museum of New Zealand Te Papa Tongarewa, PO Box 467, Wellington 6140, New Zealand *** Department of Conservation, PO Box 29, Te Anau 9600, New Zealand

ABSTRACT: The flax Anagotus fairburni (Brookes, 1932) and knobbled weevil Hadramphus stilbocarpae Kuschel, 1971 were among the first New Zealand to be granted legal protection. Both are large flightless species with narrow host– requirements. Their disjunct distributions are probably the result of predation by introduced rodents, with populations of both having apparently been extirpated by ship rats (Rattus rattus) at one documented site (Taukihepa/Big South Cape Island). Within Fiordland, flax weevils were previously known from a single small island in Breaksea Sound, and knobbled weevils had been reported from five outer islands, from Secretary Island south to Resolution Island. We report the presence of both species in Dusky Sound, and flax weevils in Chalky and Preservation Inlets, based on surveys of 134 islands in 2016 and 2017. Signs of flax weevil feeding were recorded on 56 widely scattered islands, with live or dead found on seven of these during the limited search time available. A single knobbled weevil was found at night on a small island in the Seal Islands, southwest of Anchor Island. The status of both weevil species is discussed in relation to the past and present distribution and control of rodents and stoats.

KEYWORDS: Anagotus fairburni, conservation, distribution, Fiordland, flax weevil, Hadramphus stilbocarpae, knobbled weevil, rats, Rattus norvegicus, Rattus rattus, threatened invertebrates.

Introduction In 1894, Richard Henry was employed as the curator of the newly created Resolution Island bird sanctuary, at the The large invertebrates of Fiordland, New Zealand, northern entrance to Dusky Sound (Hill & Hill 1987). During are poorly known. Most survey effort has been focused his 14 years in residence, Henry was an astute observer of on Dusky Sound, which has been a focal site for scientific wildlife, but made little mention of terrestrial invertebrates discovery and conservation management for nearly 250 years during his travels throughout Dusky and Breaksea Sounds. (Henry 1903; Beaglehole 1961; Hill & Hill 1987; Miskelly et Apart from discussing sandflies and vegetable caterpillars al. 2017a,b). It was the first site in New Zealand to receive (hepialid moth larvae mummified by Ophiocordyceps fungus), prolonged scrutiny by European naturalists, when Captain his only mention of insects and spiders was as a food source James Cook based HMS Resolution there for six weeks in for birds (Henry 1898, 1902), although he did provide March–May 1773 (Beaglehole 1961; Hoare 1982). Johann samples of caterpillars and to James Hector at the and Georg Forster and Anders Sparrman prepared extensive Colonial Museum and William Benham at the University illustrations and descriptions of , birds, seals, fishes and of Otago (Hill & Hill 1987). The focus on management of molluscs, although ‘The only insects mentioned specifically threatened birds in Dusky Sound, including vertebrate pest were the pestilential ones, the sandflies [Austrosimulium spp.] control to achieve this end, has continued to the present day and the human lice [Pediculus humanus]’ (Andrews 1986). (Elliott et al. 2010; Edge et al. 2011; Wildland Consultants Flax weevil and knobbled weevil in Fiordland 21

and Department of Conservation 2016; Department of McGuinness 2001). They are both currently ranked ‘At Conservation 2017; Miskelly et al. 2017a,b). Risk – Relict’, as they are believed to occupy less than 10% The Tamatea/Dusky Sound conservation and restoration plan of their original range (Townsend et al. 2008; Leschen et al. (Wildland Consultants and Department of Conservation 2012). Flax weevils feed only on harakeke () 2016) acknowledged that the area had not been fully explored and wharariki (P. cookianum) (Kuschel 1982; Meads 1990). for terrestrial invertebrates, but identified several features of Knobbled weevils were originally considered host-specific the invertebrate fauna considered to be distinctive (p. 44). on two species of pünui (Stilbocarpa lyallii and S. robusta) on The conservation and restoration plan covers both Dusky islands from Foveaux Strait south to the Snares Islands/Tini Sound and the adjacent Breaksea Sound, which joins the Heke (Kuschel 1971). However, populations subsequently Tasman Sea north of Resolution Island. Within this broader discovered in Fiordland were feeding on Anisotome lyallii area, some of the notable invertebrates listed were known (Meads 1990; Peat & Patrick 1996; Craw 1999). only from Breaksea Sound and the outer coast of Resolution Dusky Sound, Chalky Inlet and Preservation Inlet are Island, including the flax weevil Anagotus fairburni (Brookes, within the broad geographic range of both the flax weevil 1932) and knobbled weevil Hadramphus stilbocarpae Kuschel, and knobbled weevil, but we were unaware of any previous 1971. These large (16–22 mm long) flightless weevils were records of these species from islands within the fiords before discovered on two rat-free islets in the outer Gilbert Islands our visits in 2016 and 2017 (based on specimens in the in Breaksea Sound during preparations to eradicate Norway New Zealand Collection and a literature search). rats (Rattus norvegicus) from nearby Häwea Island and Although we were primarily engaged in surveys for petrel Breaksea Island (Meads 1990; Taylor & Thomas 1993). Both (Aves: Procellariiformes: Procellariidae) breeding colonies weevil species were translocated to Breaksea Island in 1991, (Miskelly et al. 2017b), we searched for flax weevils and three years after the rats were eradicated there (Thomas et al. knobbled weevils, or their feeding sign, on all the islands 1992; Peat & Patrick 1996; McGuinness 2001). visited. Much of the evidence we found of weevil presence The flax weevil and knobbled weevil were among the first was indirect (i.e. feeding sign), and no live specimens were New Zealand insects to receive absolute legal protection, collected of these protected species. We provide details of through their inclusion in the original Seventh Schedule of what we found as evidence of their presence, and to provide the Wildlife Act 1953, in 1980 (Miskelly 2014). Protection a baseline of distribution records for further surveys for these for the two species was initiated by their inclusion in a list two species within coastal areas of southern Fiordland. of endangered insects that ‘occur in very low numbers, have a very restricted distribution, or have only one or two surviving populations’, published in the first issue of The Methods Weta (Ramsay & Gardner 1977). This listing came soon Two boat-based surveys of islands in southern Fiordland, after reports of the vulnerability of these weevils to rodent Fiordland National Park, were undertaken during 15–24 predation (Kuschel 1971; Watt 1975). Both species were November 2016 (Dusky Sound) and 20–26 November 2017 among the large flightless beetles apparently extirpated from (mainly Chalky Inlet and Preservation Inlet). Landings were Taukihepa/Big South Cape Island, off southern Stewart made from a small inflatable dinghy, with one to nine team Island/Rakiura, during an irruption of recently colonised members landing on each island for between three minutes ship rats (Rattus rattus) there in the early 1960s (Kuschel and four hours (average 54 minutes: Tables 1 and 2). 1971, 1982; Bell 1978; Ramsay 1978). Live weevils of both A total of 134 islands was surveyed by one or more team species were last collected on Taukihepa/Big South Cape members. Few of the smaller islands had individual names Island in 1964, but only rat-gnawed elytra (of both species) on available maps and charts (where most are named as were found in 1968 (Kuschel 1971; Grace Hall, pers. comm., clusters of islands), and so we created names/numbers for 27 April 2017). them, usually numbering islands in each cluster from north The flax weevil and knobbled weevil both have to south and west to east. Central latitude and longitude extensive distributions, with the flax weevil recorded from reference points for each island are provided in Tables 1 and Manawatäwhi/Three Kings Islands south to Taukihepa/ 2; Table 1 also includes a reference number for each island Big South Cape Island (Kuschel 1982; Marris 2001), and surveyed in Dusky or Breaksea Sounds, or off Resolution the knobbled weevil from Secretary Island, Fiordland, Island, based on a GIS database of 713 islands in Dusky south to the Snares Islands/Tini Heke (Kuschel 1971; and Breaksea Sounds created by Wildland Consultants for 22 Tuhinga, Number 29 (2018)

the Department of Conservation (DOC) (see Wildland searched for feeding animals on foliage of both plant species Consultants and Department of Conservation 2016). Areas at night on three islands (one site for Anisotome, two sites for islands in Dusky Sound or near Resolution Island were for flax). obtained from the DOC Dusky + Breaksea GIS database; Details of predator control history on islands in Dusky areas for islands in Chalky Inlet and Preservation Inlet were Sound were provided by Pete McMurtrie (DOC Te estimated using ‘DOCgis’ software, by tracing the vegetation Anau) and former DOC staff member Murray Willans. edge of each island. Similar data were summarised for islands in Chalky Inlet Most islands were surveyed in daylight, and only five and Preservation Inlet by author Colin Bishop. There are islands were visited at night. We were aware of the possibility currently more than 2900 DOC 150 stoat traps set on of encountering flax weevils and knobbled weevils (both of 23 islands in Dusky Sound (2351 on Resolution Island), which feed only at night), and so searched for evidence which are checked three times per annum. Trapping effort of their presence wherever their host plants (wharariki in Chalky and Preservation Inlets was initially focused and Anisotome lyallii, respectively) were noted. Authors on removing stoats (Mustela erminea) from Chalky Island, Colin Miskelly and Alan Tennyson had encountered the the Passage Islands and Coal Island, and preventing them weevils and their feeding sign on numerous occasions re-establishing. Traps targeting two rat species and stoats previously. We have seen many hundreds of flax weevils have recently been placed on many medium-sized islands on six islands between the Poor Knights Islands and Cook in Chalky and Preservation Inlets, but the mammalian pest Strait (particularly on Mana Island), and dozens of knobbled status of most of these islands is not well known, particularly weevils on the Snares Islands/Tine Heke and also Solander for rodents. Island (Hautere) (on two species of pünui). Wharariki leaves were checked for browse damage Results on the edges of their blades, with ragged, asymmetrical Weevils and their host plants damage considered diagnostic for flax weevil presence (Fig. 1A; and see Martin 2015). This feeding sign was readily Wharariki was a minor component of coastal vegetation on distinguished from the straight-edged, symmetrical notches most islands visited, but was absent from those sites most made by Tmetolophota sp. (Noctuidae) caterpillars (Martin exposed to salt spray. In contrast, Anisotome lyallii was present 2015). Knobbled weevil browse sign was searched for on only on the most exposed, outermost islands and headlands, Anisotome stems. Where browse sign was noted, and time forming dense patches on just a few of the islands we visited permitted, we searched for live or dead animals among (e.g. ‘Centre Island’, Entry Island, ‘Northern Seal Islet’ and the plant bases and dead foliage of the affected plants. We ‘Seal Islet 1’).

A B C D

Figure 1. A. Adult flax weevil (Anagotus fairburni) on wharariki (Phormium cookianum) at night, showing characteristic flax weevil feeding sign, islet in Anchor Island Harbour, Dusky Sound/Tamatea. Photograph: Jean-Claude Stahl, Te Papa. B. Adult flax weevil on wharariki at night, Small Craft Harbour Islands (western main island), Chalky Inlet. Photograph: Colin Miskelly. C. Adult flax weevil on wharariki, Round Island, Preservation Inlet. Photograph: Colin Miskelly. D. Flax weevil larva, Round Island, Preservation Inlet. Photograph: Colin Miskelly. Flax weevil and knobbled weevil in Fiordland 23

Flax weevil feeding sign was noted on wharariki on 56 in Dusky Sound (Wildland Consultants and Department of widely scattered islands ranging in size from 0.02 ha to 514 Conservation 2016; Department of Conservation 2017). ha (Figs 2 and 3; Tables 1 and 2). Confirmation of our ability Ship rats were eradicated from Indian Island in 2010, to identify flax weevil feeding sign correctly was obtained on and single animals were caught on Resolution Island in seven islands where we had sufficient time to search for adult 2006 and 2017 (Pete McMurtrie, pers. comm., 1 September weevils, which remain well hidden during daylight. Live 2017). Elsewhere in Dusky Sound, ship rats are present on adults were found on five islands, and parts of dead adults (a Long and Cooper Islands and some adjacent islands, while head and an abdomen) on two other islands (Table 3). On Norway rat and Pacific rat or kiore (Rattus exulans) are not ‘Anchor Island Harbour Islet 1’, a single weevil was found known to be present (Pete McMurtrie, pers. comm., 27 at the base of a large wharariki plant (which had feeding March 2017). sign) after 10 minutes of searching in daylight. When the Mice were also eradicated from Indian Island in 2010, same plant was checked after dark (11pm), five adults were but were considered to be present on 11 islands we out feeding on separate leaves (Fig. 1A). Two live adults were surveyed, including six where we found flax weevil feeding found at night on the western main island of the Small Craft sign (mouse records from Pete McMurtrie, pers. comm., Harbour Islands, Chalky Inlet (Fig. 1B), and two adults and 27 March 2017 based on incidental catches in stoat traps; five larvae (the latter among wharariki roots) were found on Table 4). Round Island in Preservation Inlet (Figs 1C and 1D). Stoats were present throughout Dusky Sound from Knobbled weevil feeding sign was recorded on ‘Seal Islet c. 1900 to c. 2000, and are likely to have visited all islands 1’ in Dusky Sound, where a single adult weevil was found other than ‘Centre Island’ (Miskelly et al. 2017b). Stoats feeding on Anisotome during a night-time visit (Figs 4 and were eradicated from Anchor Island in 2001, Parrot Island 5). South of Dusky Sound, possible knobbled weevil feeding in 2005 and Indian Island in 2010, and have been controlled sign was also noted on the largest Zero Nugget islet, off the to low density on Resolution Island since 2008. Other than north coast of Chalky Island. Resolution Island, none of the islands we surveyed in Dusky Sound currently support resident stoats. However, single stoats were caught on ‘Shag Island 3’ in May 2007, and on Rodents and mustelids Front Island and ‘Shag Island 6’ in July 2014 (Miskelly et al. With the exception of mice (Mus musculus) (on 11 islands), 2017b). along with stoats and red deer (Cervus elaphus) (both Rodents have never been known to occur on 514 ha on Resolution Island only), other species of introduced Chalky Island; stoats were eradicated here and on the nearby mammals were considered absent from the islands surveyed Passage Islands in 1999 (Department of Conservation 2017).

Table 1 [p 24-25] Island locations, search effort and evidence for the presence of flax weevils and knobbled weevils on islands in Dusky Sound, November 2016, and off the west coast of Resolution Island and the entrance to Breaksea Sound in November 2017. ID ISLAND is the Department of Conservation (DOC) Dusky + Breaksea Sound island database reference number for each island. Island names within inverted commas are tag names derived from nearby named geographical features or island group names. ‘Weevils’ refers to whether evidence of flax weevils or knobbled weevils was recorded during the surveys: Live = 1 or more live weevils of either species found; Dead = remains of a dead flax weevil found; Sign = flax weevil feeding sign observed without live or dead weevils being seen. ‘Stoat trap’ refers to whether at least one stoat trap was maintained on the island by DOC at the time of our visit. Observers: AS = Andrew Smart; AT = Alan Tennyson; CB = Colin Bishop; CM = Colin Miskelly; DA = Dougal Austin; GT = Graeme Taylor; HE = Hannah Edmonds; J-CS = Jean-Claude Stahl; LM = Lawrie Mead; RP = Riki Parata; RW = Rick Webber; TG = Terry Greene. ‘Time’ is the approximate length of time (hours:minutes) that observers were ashore. 24 Tuhinga, Number 29 (2018) 5 25 15 50 35 25 50 40 25 30 30 40 30 35 55 30 20 15 30 20 50 25 1:50 2:25 1:45 1:50 2:05 1:30 1:30 1:45 1:00 4:00 Time Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Night Day + night Day Day/night AS AS CM CB, LM CB, AT, CM AT, CM AT, CM CM & AS CM & AS AT & CM AT & CM AT & CM AT & CM AT & CM AT & CM AT & CM AT & CM AT AT & CM AT AS & J-CS AS & J-CS AS & J-CS AT & J-CS AT & J-CS AT Observers AT, CM & HE AT, AT, CM & J-CS AT, CM & J-CS AT, CM & J-CS AT, CM & J-CS AT, AT, CM, J-CS & AS J-CS & CM, AT, AT, CM, HE & RW CM, AT, AT, CM, J-CS & HE CM, AT, AT, CB, CM, LM, RP LM, CM, CB, AT, Date 26 Nov 17 26 Nov 17 25 Nov 17 25 Nov 17 25 Nov 17 25 Nov 16 20 Nov 16 20 Nov 16 18 Nov 16 18 Nov 16 18 Nov 16 18 Nov 16 22 Nov 16 18 Nov 16 18 Nov 16 18 Nov 16 22 Nov 16 23 Nov 16 22 Nov 16 23 Nov 16 24 Nov 16 17 Nov 16 18 Nov 16 17 Nov 16 17 Nov 16 17 Nov 16 17 Nov 16 17 Nov 16 17 Nov 16 17 Nov 16 22 Nov 16 22 Nov 17 & 24 Nov 16 17 & 24 Nov Ye s Ye s Ye s Ye s Ye s Ye s Ye s Ye s Ye s Ye s Ye s Ye s No No No No No No No No No No No No No No No No No No No No Stoat trap No No No No No No No No No No No No No No Live Live Sign Sign Sign Sign Sign Sign Sign Sign Sign Sign Sign Sign Sign Sign Dead Dead Weevils 6666˚ 6089˚ 5292˚ 4991˚ 4481˚ 7183˚ 7180˚ 6300˚ 5524˚ 5442˚ 5344˚ 5387˚ 5605˚ 5437˚ 5451˚ 5454˚ 5206˚ 5227˚ 5148˚ 5078˚ 5180˚ 5272˚ 5279˚ 5256˚ 5365˚ 5428˚ 5538˚ 5584˚ 5613˚ 5661˚ 5791˚ 5812˚ ...... 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 Longitude E 5567˚ 6035˚ 6387˚ 7447˚ 7435˚ 6547˚ 6573˚ 6700˚ 6885˚ 6982˚ 7075˚ 7100˚ 7083˚ 7094˚ 7083˚ 7250˚ 7353˚ 7376˚ 7413˚ 7455˚ 7594˚ 7467˚ 7471˚ 7530˚ 7378˚ 7384˚ 7419˚ 7415˚ 7437˚ 7417˚ 7509˚ 7485˚ ...... 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 Latitude S Latitude 6 . 6 2 8 1 3 9 5 9 7 9 9 3 6 6 4 1 5 5 6 3 1 1 0 0 6 3 6 3 2 2 ...... 0 0 2 0 1 0 0 1 3 0 2 2 0 0 0 1 3 0 6 4 1 0 5 0 0 3 13 40 72 21 20887 1136 Area (ha) Area 4 3 28 66 65 64 26 21 25 699 661 619 353 361 586 583 530 511 484 485 491 444 363 360 337 400 397 386 391 388 389 368 ID ISLAND Table 1 Table Island name ‘Roof-top stack’ Resolution islet’ ‘North-west stack’ Finger Peninsula ‘Nth Five Fingers) (Five ‘Outer Finger’ Fingers) (Five ‘Inner Finger’ Islet 1’ ‘Acheron Islet 2’ ‘Acheron Resolution Island end) (west Islet’ ‘ Cove Islet 1’ ‘Goose Cove Island Parrot Islet 1’ ‘Parrot Pigeon Island ‘Pigeon Islet 1’ ‘Pigeon Islet 2’ Island’ ‘Centre Island Petrel ‘North’ Islet’ ‘Petrel end) Island (west Petrel ‘Main’ Entry Island Anchor Island (nth headland) ‘Anchor Island Harbour Islet 1’ ‘Anchor Island Harbour Islet 2’ ‘Anchor Island Harbour Islet 3’ Anchor Islet 1’ ‘North-east Anchor Islet 2’ ‘North-east Anchor Islet 3’ ‘North-east Anchor Islet 4’ ‘North-east Anchor Islet 5’ ‘North-east Anchor Islet 6’ ‘North-east Useless Island ‘Useless Islet 1’ Flax weevil and knobbled weevil in Fiordland 25 25 50 30 20 20 50 20 15 20 20 30 35 35 10 35 45 35 30 1:10 1:25 1:15 1:15 3:10 1:00 1:20 2:00 1:25 1:00 4:00 3:00 2:40 1:00 Time Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Night Day/night AS AT AT AT AT AT CM CM CM J-CS CM & AS CM & AS AT & CM AT & CM AT & CM AT AT & CM AT & CM AT AT & J-CS AT J-CS & DA J-CS & DA J-CS & DA J-CS & HE Observers AT, CM & AS CM & AT, AT, CM & AS CM & AT, AT, CB & CM AT, AT, CM, HE & RW CM, AT, AT, CM, J-CS & AS J-CS & CM, AT, AT, CM, J-CS & HE CM, AT, J-CS, HE, RW & DA RW HE, J-CS, AT & CM; GT & TG GT & & CM; AT AT, CM, J-CS, AS & DA J-CS, CM, AT, AS & DA J-CS, CM, AT, Date 20 Nov 16 20 Nov 16 19 Nov 16 19 Nov 16 19 Nov 16 19 Nov 16 19 Nov 16 19 Nov 16 19 Nov 16 19 Nov 16 19 Nov 16 Nov 16 16 Nov 16 16 Nov 16 Nov 16 16 Nov 21 Nov 16 21 Nov 16 23 Nov 16 23 Nov 16 23 Nov 16 21 Nov 16 20 Nov 16 20 Nov 16 20 Nov 16 21 Nov 16 21 Nov 16 21 Nov 17 25 Nov 21 Nov 16 21 Nov 16 23 Nov 23 Nov 16 23 Nov 22 Nov 16 22 Nov 16 23 Nov 16 15 Nov 23 Nov 16 & 25 Nov 17 16 & 25 Nov 23 Nov Ye s Ye s Ye s Ye s Ye s Ye s Ye s Ye s Ye s Ye s Ye s Ye s No No No No No No No No No No No No No No No No No No No No Stoat trap No No No No No No No No No No No No No No No No No Live Live Sign Sign Sign Sign Sign Sign Sign Sign Sign Sign Sign Sign Sign Weevils 5897˚ 7256˚ 7281˚ 7680˚ 7697˚ 7725˚ 7746˚ 7751˚ 7836˚ 7884˚ 5541˚ 5688˚ 5441˚ 4828˚ 4918˚ 4931˚ 4839˚ 4998˚ 4978˚ 5016˚ 5000˚ 4994˚ 5022˚ 5042˚ 4981˚ 4952˚ 4836˚ 4822˚ 5846˚ 4796˚ 4776˚ 4796˚ ...... 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 Longitude E 7660˚ 7780˚ 7404˚ 7403˚ 7321˚ 7318˚ 7322˚ 7324˚ 7330˚ 7343˚ 7359˚ 7774˚ 7765˚ 7757˚ 7819˚ 7753˚ 7780˚ 7775˚ 7794˚ 7785˚ 7781˚ 7763˚ 7994˚ 8011˚ 7776˚ 7803˚ 7801˚ 7784˚ 7528˚ 7749˚ 7809˚ 7780˚ ...... 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 Latitude S Latitude 6 4 1 7 . 0 1 2 2 1 2 6 3 3 1 6 2 3 5 1 1 8 3 2 2 3 6 3 5 7 5 8 . . . 04 ...... 0 1 0 0 0 0 0 1 0 4 0 0 7 1 2 2 0 0 0 1 0 0 0 1 9 1 0 16 20 0 13 167 Area (ha) Area 2 6 32 36 37 31 14 71 41 47 49 56 51 53 54 106 105 458 454 290 295 308 286 292 276 280 282 294 257 253 351 29+34 ID ISLAND (continued) Indian Island (north coast) Island Front Island’ ‘Little Front ‘Shag Island 1’ ‘Shag Island 2’ ‘Shag Island 3’ ‘Shag Island 4’ ‘Shag Island 5’ ‘Shag Island 6’ ‘Cooper Islet’ Thrum Cap Island (south coast) Passage ‘East Seal Islet 1’ ‘East Seal Islet 2’ ‘Southern Seal Islet’ Island 1’ ‘Many Island 2’ ‘Many Island 3’ ‘Many Island 4’ ‘Many Islet 1’ ‘Many Islet 2’ ‘Many Islet 3’ ‘Many ‘South Dusky Islet’ ‘South Dusky Stack’ Nomans Island (east end) ‘Seal Islet 2’ ‘Seal Islet 3’ ‘Useless Islet 2’ ‘Northern Seal Islet’ ‘Seal Island west’ ‘Seal Islet 1’ ‘Seal Island east’ Table 1 Table Island name 26 Tuhinga, Number 29 (2018)

Fig. 2 Localities with evidence of flax weevil presence on islands in Dusky Sound and Breaksea Sound in November 2016 and November 2017. Circles = feeding sign only; squares = feeding sign plus live or dead animals; crosses = islands visited without flax weevil feeding sign being recorded. The numbers 31, 363 and 444 are the Department of Conservation Dusky + Breaksea Sound island database reference numbers for three unnamed islands where live flax weevils were found (see Table 3). Within Fiordland, flax weevils were known previously only from Wairaki Island in Breaksea Sound. Knobbled weevils were known from OG3 (= Outer Gilbert Island 3), Wairaki Island, and the outer coast of Resolution Island, including Five Fingers Peninsula. Both species were introduced to Breaksea Island in 1991. Further detail of the islands southwest of Anchor Island (inset) is given in Fig. 4. Flax weevil and knobbled weevil in Fiordland 27

Fig. 3 Localities with evidence of flax weevil presence on islands in Chalky Inlet and Preservation Inlet and connected waterbodies (Cunaris Sound, Isthmus Sound and Long Sound) in November 2017. Circles = feeding sign only; squares = feeding sign plus live animals; crosses = islands visited without flax weevil feeding sign being recorded.

Table 2 [p 28-29] Island locations, search effort and evidence for the presence of flax weevils and knobbled weevils on islands in Chalky Inlet and Preservation Inlet and their inland extensions, November 2017. ‘Waterbody’ refers to whether the island is located in Chalky Inlet, Cunaris Sound, Preservation Inlet, Isthmus Sound or Long Sound. Island names within inverted commas are tag names derived from nearby named geographical features, island group names, or notable species found during the survey. ‘Weevils’ refers to whether evidence of flax weevils or knobbled weevils was recorded during the surveys: Live = 1 or more live weevils of either species found; Dead = remains of a dead flax weevil found; Sign = flax weevil feeding sign observed without live or dead weevils being seen. ‘Trap’ refers to whether at least one stoat trap (DOC 150) or resetting A24 trap was maintained on the island at the time of our visit. Observers: AT = Alan Tennyson; CB = Colin Bishop; CM = Colin Miskelly; CP = Chris Pascoe; GT = Graeme Taylor; LM = Lawrie Mead; PK = Pete Kirkman; RP = Riki Parata; TG = Terry Greene. ‘Time’ is the approximate length of time (hours:minutes) that observers were ashore. 28 Tuhinga, Number 29 (2018) 3 3 40 15 19 10 47 15 15 12 32 25 26 21 15 56 10 48 30 52 13 55 20 30 15 1:26 1:05 1:10 2:20 3:45 2:00 1:36 1:05 1:15 3:15 Time Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day + + night Day Day + + night Day Day/Night AT CM CM CM CM CM RP & TG AT & GT AT AT & LM AT AT & LM AT & LM AT CB & LM CB & LM CB & LM CB & LM CB & LM AT & CM AT & CM AT & CM AT CM & TG Observers AT, RP & TG RP & AT, AT, GT & RP AT, CB, LM & TG LM & CB, CB, LM & RP CB, AT, CM & LM AT, CM, RP & TG RP & CM, GT, LM & RP GT, AT, CB, GT & TG GT & CB, AT, CB, LM, RP & TG RP & LM, CB, AT, CM, GT, LM & RP GT, CM, AT, AT, CB, CM, GT, LM, RP & TG RP & LM, GT, CM, CB, AT, TG RP & LM, GT, CM, CB, AT, AT, CB, CM, GT, LM, RP & TG RP & LM, GT, CM, CB, AT, TG RP & LM, GT, CM, CB, AT, AT, CB, CM, GT, LM, RP & TG RP & LM, GT, CM, CB, AT, Date 23 Nov 17 23 Nov 17 23 Nov 17 23 Nov 17 24 Nov 17 24 Nov 17 24 Nov 17 25 Nov 17 24 Nov 17 25 Nov 17 25 Nov 23 Nov 17 23 Nov 17 23 Nov 17 23 Nov 17 24 Nov 17 23 Nov 17 23 Nov 17 23 Nov 17 23 Nov 17 23 Nov 17 22 Nov 17 22 Nov 17 22 Nov 17 22 Nov 17 24 Nov 17 24 Nov 17 25 Nov 17 25 Nov 17 25 Nov 17 24 Nov 17 24 Nov 17 25 Nov 17 25 Nov 24 Nov 17 24 Nov 23 Nov 17 23 Nov 17 25 Nov Ye s Ye s Ye s Ye s Ye s Ye s Ye s No No No No No No No No No No No No No No No No No No No No No No No No No No Ye s Ye s Trap No No No No No No No No No No No No No No No No No No No Live Sign Sign Sign Sign Sign Sign Sign Sign Sign Sign Sign Sign Sign Sign Sign Weevils 5233˚ 6564˚ 5142˚ 5154˚ 5186˚ 5194˚ 5386˚ 5462˚ 5425˚ 5413˚ 5364˚ 5349˚ 5297˚ 5429˚ 5693˚ 5702˚ 5738˚ 5754˚ 5900˚ 5905˚ 5934˚ 6005˚ 6024˚ 5932˚ 5891˚ 5949˚ 5963˚ 6412˚ 6483˚ 6441˚ 6395˚ 6474˚ 6484˚ 6569˚ 6580˚ ...... 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 Longitude E 0352˚ 0351˚ 0356˚ 0345˚ 0323˚ 0330˚ 0328˚ 0310˚ 0309˚ 0276˚ 0161˚ 0846˚ 0684˚ 0654˚ 0656˚ 0287˚ 0310˚ 0327˚ 0221˚ 0179˚ 9719˚ 9722˚ 9714˚ 9716˚ 9674˚ 9686˚ 9702˚ 9704˚ 9727˚ 9725˚ 9648˚ 9713˚ 0491˚ 9663˚ 0238 ˚ ...... 46 46 46 46 46 46 46 46 46 46 46 46 45 45 45 45 46 46 46 46 46 46 46 46 45 45 45 45 45 45 45 45 46 45 46 Latitude S 2 2 2 1 2 7 2 1 0 7 1 4 7 1 1 1 1 . . 12 03 06 04 02 06 05 02 05 02 01 07 01 03 03 06 ...... 0 0 8 0 0 1 5 0 0 0 0 0 0 0 2 168 514 0 0 0 0 0 0 0 0 0 0 0 0 0 18 24 0 0 0 Area Chalky Chalky Chalky Chalky Chalky Chalky Chalky Chalky Chalky Chalky Chalky Chalky Chalky Chalky Chalky Chalky Chalky Chalky Chalky Chalky Chalky Chalky Chalky Chalky Chalky Chalky Chalky Chalky Chalky Chalky Chalky Chalky Chalky Chalky Chalky Waterbody Island name Chalky Island Nugget (‘main’) Zero Nugget (‘eastern’)Zero Finger Rock (‘inner’) Finger Rock (‘eastern tall stack’) Island (‘south or main’) Passage ‘sth coast stack 1’ Island Passage ‘sth coast stack 3’ Island Passage ‘sth coast stack 4’ Island Passage ‘sth coast stack 5’ Island Passage ‘sth coast stack 6’ Island Passage ‘sth coast stack 7’ Island Passage Island (‘north’) Passage ‘Stack NW of Gulches Head’ ‘Stack NE of Red Head 1’ ‘Stack NE of Red Head 2’ ‘Stack NE of Red Head 3’ Island (‘outer or little’) Garden Island’ ‘Islet NW of main Garden Island (‘main’) Garden ‘Islet south of Stripe Head 1’ ‘Islet south of Stripe Head 2’ Little Island – northern peninsula islet’) Little Island (‘west islet’) Little Island (‘middle tiny Little Island (‘east islet’) Small Craft Hrbr Is (‘main west’) Small Craft Hrbr Is (‘main east’) islet’) Small Craft Hbr Is (‘korure Small Craft Hbr Is (‘SW islet’) Small Craft Hbr Is (‘SE islet 1’) Small Craft Hbr Is (‘SE islet 2’) Small Craft Hbr Is (‘NE island’) Small Craft Hbr Is (‘NE islet’) Small Craft Hbr Is (‘mid-channel’) Table 2 Table Flax weevil and knobbled weevil in Fiordland 29 7 15 55 30 40 20 10 20 16 15 31 54 35 30 21 53 30 20 30 20 40 15 10 48 30 15 12 1:30 1:04 2:00 2:05 1:15 1:10 2:04 1:35 Time Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day Day/Night AT AT CB TG CM CB & TG AT & CM AT & CM AT CM & TG CM & TG CB & GT Observers CB, LM & TG LM & CB, TG LM & CB, AT, CM & TG CM & AT, AT, CM & TG CM & AT, TG CM & AT, CB, CM & TG CM & CB, TG CM & CB, CB, CM & TG CM & CB, TG CM & CB, TG CM & CB, AT, GT, LM & RP GT, AT, CB, GT, LM & RP GT, CB, AT, CM, GT & RP CM, AT, AT, GT, LM & RP GT, AT, CB, GT, LM, RP & TG RP & LM, GT, CB, AT, CM, LM, RP & TG RP & LM, CM, AT, AT, CB, CM, GT, LM, RP & TG RP & LM, GT, CM, CB, AT, TG RP & LM, GT, CM, CB, AT, TG RP & LM, GT, CM, CB, AT, TG RP & LM, GT, CM, CB, AT, TG RP & LM, GT, CM, CB, AT, TG RP & LM, GT, CM, CB, AT, TG RP & LM, GT, CM, CB, AT, AT, CB, CM, CP, GT, LM, PK, RP & TG RP & PK, LM, GT, CP, CM, CB, AT, Date 25 Nov 17 25 Nov 17 25 Nov 17 25 Nov 17 22 Nov 17 22 Nov 17 21 Nov 17 20 Nov 25 Nov 17 25 Nov 17 22 Nov 17 22 Nov 17 22 Nov 17 22 Nov 17 22 Nov 17 22 Nov 17 20 Nov 17 21 Nov 17 21 Nov 17 21 Nov 17 21 Nov 17 21 Nov 17 21 Nov 17 21 Nov 17 21 Nov 17 21 Nov 17 21 Nov 17 21 Nov 17 21 Nov 17 20 Nov 17 20 Nov 20 Nov 17 20 Nov 17 20 Nov 17 20 Nov 17 20 Nov 17 21 Nov 17 21 Nov Ye s Ye s Ye s Ye s Ye s Ye s Ye s Ye s Ye s No No No No No No No No No No No No No No No No No No No No No No No No No No Trap No No No No No No No No No No No No No No No No No No No No No No No No No No Live Sign Sign Sign Sign Sign Sign Sign Sign Weevils 7340˚ 7406˚ 7395˚ 7400˚ 8716˚ 8764˚ 6566˚ 6653˚ 6661˚ 7508˚ 5686˚ 6296˚ 6457˚ 6400˚ 6391˚ 6651˚ 6826˚ 6961˚ 7099˚ 6630˚ 6651˚ 6670˚ 6796˚ 6814˚ 6804˚ 6912˚ 6928˚ 6932˚ 6918˚ 6939˚ 6623˚ 6902˚ 7535˚ 7479˚ 7517˚ ...... 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 166 Longitude E 9758˚ 9734˚ 9732˚ 9620˚ 0904˚ 0722˚ 0830˚ 0786˚ 0767˚ 0973˚ 0869˚ 0949˚ 1037˚ 0793˚ 0788˚ 0771˚ 0737˚ 0701˚ 0690˚ 0668˚ 0680˚ 0702˚ 0716˚ 0753˚ 0656˚ 0525˚ 0260˚ 0204˚ 0259˚ 0153˚ 0146˚ 0170˚ 0170˚ 9603˚ 9600˚ ...... 46 46 46 46 46 46 46 46 45 46 46 46 45 45 45 46 46 46 46 46 46 46 46 46 46 46 46 46 46 45 46 46 46 46 45 Latitude S 7 5 6 . 7 2 7 4 6 3 5 1 2 1 1 9 1 0 1 6 3 3 6 4 8 3 8 4 6 . . 03 01 03 05 03 01 ...... 13 0 0 2 0 0 0 2 0 2 4 0 3 0 0 0 1 0 0 0 0 0 3 0 1 0 0 0 0 0 57 0 0 24 Area 109 Long Long Long Long Long Long Long Long Long Isthmus Cunaris Cunaris Cunaris Cunaris Preservation Preservation Preservation Preservation Preservation Preservation Preservation Preservation Preservation Preservation Preservation Preservation Preservation Preservation Preservation Preservation Preservation Preservation Preservation Preservation Preservation Waterbody (continued) ‘Lady Bay northeast islet’ ‘Lady Bay Only Island ‘Only islet’ Island name ‘Cunaris SW islet (outer)’ ‘Cunaris SW islet (middle)’ ‘Cunaris SW islet (inner)’ islet’ ‘Cliff Cove ‘Gulches Head stack’ Spit Island Head islet’ ‘Cavern Head’ of Cavern ‘Stack west Head’ ‘Stack NW of Cavern Steep-to Island Round Island or Long Island Weka Cemetery Island Is (‘outermostCording islet’) Is (‘outer’) Cording northern islets’) Is (‘outer, Cording Is (‘main’) Cording Is (‘inner NW islet’) Cording Is (‘outer NW islet’) Cording Is (‘NE 28 m asl’) Cording islet SE of NE 28 m asl’ ‘Cording Is (‘NE 35 m asl’) Cording islet SW of NE 35 m asl’ ‘Cording B & C’) A, Is (‘east islet Cording Single Tree Island Sound islet’ ‘Isthmus south – sthnmost islet’ Cove ‘Jane south – main islet’ Cove ‘Jane islet’ sth – middle tiny Cove ‘Jane main islet’ ‘Lady Bay islet’ SE inner tiny ‘Lady Bay islet’ SE outer tiny ‘Lady Bay Table 2 Table 30 Tuhinga, Number 29 (2018)

Discussion rat predation was the likely reason broad-billed prions (Pachyptila vittata) disappeared from Anchor Island before The main finding of these surveys was that flax weevils were the arrival of stoats in Dusky Sound in 1900. Rats are widespread and apparently common on islands in southern unlikely to have reached many of the other islands we Fiordland. Evidence of their presence was found on 56 surveyed in Dusky Sound, as few of them provide safe widely spread islands, with live or dead animals confirmed anchorages and most are beyond the swimming range on seven of these islands. Before the 2016 and 2017 surveys, of rats (Atkinson 1986; Russell et al. 2005). Norway rats only a single natural flax weevil population was known in disappeared rapidly from Resolution Island after the arrival Fiordland, on Wairaki Island in Breaksea Sound (Thomas et of stoats, and they remain absent (Taylor 1978; Wildland al. 1992; Peat & Patrick 1996; McGuinness 2001). Consultants and Department of Conservation 2016). We found only a single knobbled weevil, but expect The rapid extirpation of rats from several large islands in that more thorough searches (especially at night) will reveal Dusky Sound after stoats arrived supports the contention them to be present on other islands with Anisotome lyallii that they were Norway rats, as this large burrowing species in southern Fiordland. is particularly vulnerable to stoat predation compared The presence of both these weevil species on islands in to the ship rat (Thomson 1921; Taylor 1978). Ship rats southern Fiordland is primarily driven by the distribution are present on islands in the southeast of Dusky Sound of their host plants, overlain by the impacts of introduced (Pete McMurtrie, pers. comm., 27 March 2017); they predators. Large flightless weevils are considered particularly were eradicated from Indian Island in 2010, but were not vulnerable to rat predation, and the presence of flax weevils considered to have been present on any of the Dusky or knobbled weevils on islands is a good indicator that rats Sound islands we visited at the time of the survey. are absent (Ramsay 1978; Bremner et al. 1984; Meads 1990; Flax weevils can apparently persist in the presence of Craw 1999). mice, based on records of flax weevils in alpine areas of Norway rats were abundant around Dusky Sound in the Tararua Ranges and Marlborough (McGuinness 2001), 1884 (Reischek 1888), and are likely to have been the along with our discovery of flax weevil feeding sign on at species of rat that Richard Henry encountered on Pigeon, least six islands where mice were believed to be present Resolution and Indian Islands in the 1890s (Henry 1895, (and on Indian Island, where ship rats and mice were 1901; Taylor 1978; Hill & Hill 1987). Although there is no eradicated in 2010). These six islands (listed in Table 4) direct evidence of rats on other islands in Dusky Sound, should be a priority for survey to confirm the presence Medway (2011) and Miskelly et al. (2017b) argued that and co-occurrence of flax weevils and mice.

Table 3 Details of flax weevils found on seven islands in southern Fiordland in 2016 and 2017. ID ISLAND is the Department of Conservation Dusky + Breaksea Sound island database reference number for each island. See Tables 1 and 2 for island locations and time spent ashore.

Island name Waterbody ID ISLAND Area (ha) No. of flax weevils seen Pigeon Island Dusky Sound 3 72.8 1 head ‘Centre Island’ Dusky Sound 444 2.1 1 live adult Entry Island Dusky Sound 26 0.6 1 abdomen ‘Anchor I Harbour Islet 1’ Dusky Sound 363 0.3 5 live adults ‘Shag Island 5’ Dusky Sound 31 0.6 2 live adults Small Craft Harbour Is – Chalky Inlet – 18 2 2 live adults ‘main west’ . Round Island Preservation Inlet – 2.6 2 live adults & 5 larvae Flax weevil and knobbled weevil in Fiordland 31

Fig. 4 Localities with flax weevil feeding sign and one live knobbled weevil on islands southwest of Anchor Island, Dusky Sound, in November 2016. Circles = flax weevil feeding sign; solid square = knobbled weevil (photographed; Fig. 5); crosses = islands visited without evidence of flax weevil being recorded. The two- or three-digit numbers are the Department of Conservation Dusky + Breaksea Sound island database reference numbers for the 16 unnamed islands that were surveyed in this portion of Dusky Sound (see Table 1). None of the islands in the main cluster of the Many Islands (on the eastern edge of the map) was included in the survey.

Stoats are competent swimmers and are likely to have reached more islands in southern Fiordland than rats (Elliott et al. 2010; Veale et al. 2012). While this would have exposed further weevil populations to predation by introduced mammals, stoats are far less damaging to invertebrates than are rats (Bremner et al. 1984). Stoats do eat insects (41% of guts from stoats collected from New Zealand national parks contained insect remains), but no weevil remains were identified in 1514 New Zealand stoat guts examined, including 451 from Fiordland (King & Moody 1982). The largest island on which we found evidence of flax weevils, the 514 ha Chalky Island in Chalky Inlet, supported stoats until they were eradicated Fig. 5 Adult knobbled weevil (Hadramphus stilbocarpae) in 1999, but no rodents have ever been recorded here on Anisotome lyallii at night, on islet 53 in the Seal (Department of Conservation 2017). Islands, Dusky Sound (Photo: Colin Miskelly).

32 Tuhinga, Number 29 (2018)

Most of the islands on which we found flax weevils were extirpated by stoats on Pigeon Island, and the six are too small to sustain resident stoat populations, and years since ship rats were eradicated on Indian Island. are sufficiently far offshore that stoats are likely to be Possible ways that flax weevils could move between only infrequent visitors (Elliott et al. 2010; Miskelly et al. islands include hitching rides on or with people, and 2017b). It follows that predation pressure by rats and stoats rafting on clumps of floating flax. Wharariki often grows on most Dusky Sound islands where we found weevils overhanging the vertical shores on islands in southern will have been negligible or non-existent. Exceptions to Fiordland, and clumps are commonly seen floating in the this were Indian Island (168 ha) and Pigeon Island (73 southern fiords after storms (Pete Young, pers. comm., 27 ha, and within 100 m of Resolution Island), where rats April 2017). and stoats were formerly present (Henry 1895, 1901; On Indian Island, we traversed 1.7 km along the north DOC trapping records provided by Pete McMurtrie, pers. coast, and the only location where flax weevil sign was comm., 27 March 2017 and Murray Willans, pers. comm., noted was at the much-visited historic site at the northeast 3 May 2017). The locations where we found flax weevil point (Peat 2007), suggesting that recent human-mediated feeding sign (and one dead weevil) on these two islands arrival is likely. The apparent distribution of flax weevils are among the most visited historic sites in Dusky Sound: on dozens of islands throughout southern Fiordland with the 1773 ‘first encounter’ site between Mäori and Captain differing predator histories makes this an ideal area to study Cook on Indian Island, and Richard Henry’s house site the dispersal ability of these flightless insects, and their on Pigeon Island (Peat 2007). Either flax weevils survived vulnerability to mammal predation, by investigating the in the presence of rats on Pigeon and Indian Islands, and genetic structure of the metapopulation. have since recovered in numbers, or they arrived from It is surprising that evidence of these weevils has not elsewhere during the 100-plus years since Norway rats been reported from Dusky Sound previously. Not only

Table 4 Twelve islands in Dusky Sound visited in November 2016 that were believed to have (or have had) mice, with details of whether flax weevils or their feeding sign was found on each island. ID ISLAND is the Department of Conservation Dusky + Breaksea Sound island database reference number for each island. See Table 1 for island locations and time spent ashore.

Island name ID ISLAND Area (ha) Flax weevil evidence Mouse status Resolution Island 28 20887 – Present ‘Cormorant Cove Islet’ 530 13.6 – Present ‘Goose Cove Islet 1’ 511 1.3 – Present Pigeon Island 3 72.8 Sign + dead Present ‘Pigeon Islet 2’ 491 2.4 – Present ‘North-east Anchor Islet 6’ 389 1.3 – Present Useless Island 25 3.2 Sign Present ‘Useless Islet 1’ 368 0.2 Sign Present ‘Useless Islet 2’ 351 0.5 Sign Present ‘Acheron Islet 1’* 586 0.9 Sign Present ‘Acheron Islet 2’ 583 0.9 Sign Present Indian Island (nth coast)† 2 167.6 Sign Eradicated 2010

* Rodent-gnawed seed noted on 20 November 2016 † Ship rats also eradicated in 2010 Flax weevil and knobbled weevil in Fiordland 33

was flax weevil feeding sign evident at two sites frequently nature reserves. Wellington: Department of Lands and visited by tourists in November 2016, but 13 of the islands Survey. 237 pp. on which we found weevils or their sign are equipped Bremner, A.G., Butcher, C.F. and Patterson, G.B. (1984). with stoat traps (Table 1), meaning that they are visited The density of indigenous invertebrates on three islands at least three times a year by DOC staff or volunteers. in Breaksea Sound, Fiordland, in relation to the With greater awareness of their characteristic feeding distribution of introduced mammals. Journal of the Royal sign, followed by targeted night-time searches, flax weevil Society of New Zealand 14: 379–386. and knobbled weevil will undoubtedly prove to be more Craw, R.C. (1999). Molytini (Insecta: Coleoptera: widely distributed in Fiordland than current knowledge Curculionidae: Molytinae). Fauna of New Zealand 39. suggests. Lincoln: Manaaki Whenua Press. 68 pp. Department of Conservation (2017). Conserving Fiordland’s Acknowledgements biodiversity 1987–2015: the challenges, the achievements, the knowledge. Ti Tiaki i te Taiao ki Tu Rua o te moko: The 2016 survey was funded by the Museum of New ngä wero, ngä haumäuiui, ngä mätauranga. Te Anau: Zealand Te Papa Tongarewa (Te Papa), Wellington, New Department of Conservation. 128 pp. Zealand, and the 2017 survey was funded by the Department Edge, K.-A., Crouchley, D., McMurtrie, P., Willans, M.J. of Conservation (DOC), New Zealand. Both surveys were and Byrom, A. (2011). Eradicating stoats (Mustela based on the DOC vessel Southern Winds. We thank erminea) and red deer (Cervus elaphus) off islands in Andrew Smart, Hannah Edmonds, Graeme Taylor and Terry Fiordland. Pp. 166–171. In: Veitch, C.R., Clout, M.N. Greene of DOC; Jean-Claude Stahl, Dougal Austin and and Towns, D.R. (eds). Island invasives: eradication and Rick Webber of Te Papa; and Riki Parata and Lawrie Mead management. Gland, Switzerland: IUCN. 542 pp. for their assistance with the surveys. Southern Winds crew Elliott, G., Willans, M., Edmonds, H. and Crouchley, D. Chris Pascoe, Pete Young and Pete Kirkman are gratefully (2010). Stoat invasion, eradication and reinvasion of thanked for their skill in getting us safely on and off the islands in Fiordland. New Zealand Journal of Zoology 37: islands, and for looking after us so well when on board. Grace 1–12. Hall (Landcare Research, New Zealand), Pete McMurtrie Henry, R. (1895). Resolution Island. Appendix to the (DOC), Murray Willans and Pete Young kindly provided Journals of the House of Representatives C-01: 97–102. information quoted here, and Susan Waugh (Te Papa) assisted Henry, R. (1898). Notes on bird-life in the West Coast with map preparation. The manuscript was improved by sounds. Transactions and Proceedings of the New Zealand comments received from Eric Edwards and John Marris. Institute 30: 279–293. Henry, R. (1901). Resolution Island. Appendix to the References Journals of the House of Representatives C-01: 132–136. Andrews, J.R.H. (1986). The southern ark; zoological discovery Henry, R. (1902). Natural history notes from Dusky Sound. in New Zealand 1769–1900. Auckland: Century Transactions and Proceedings of the New Zealand Institute Hutchinson. 237 pp. 34: 570–571. Atkinson, I.A.E. (1986). Rodents on New Zealand’s northern Henry, R. (1903). The habits of the flightless birds of New offshore islands: distribution, effects and precautions Zealand; with notes on other New Zealand birds. against further spread. Pp. 13–40. In: Wright, A.E. and Wellington: Government Printer. 88 pp. Beever, R.E. (eds). The offshore islands of northern New Hill, S. and Hill, J. (1987). Richard Henry of Resolution Zealand. Wellington: Department of Lands and Survey. Island. Dunedin: John McIndoe in association with the 255 pp. New Zealand Wildlife Service. 364 pp. Beaglehole, J.C. (ed.) (1961). The journals of Captain James Hoare, M.E. (ed.) (1982). The Resolution journal of Johann Cook on his voyages of discovery. II. The voyage of the Reinhold Forster 1772–1774. Vol. 2. London: Hakluyt Resolution and Adventure 1772–1775. Hakluyt Society Society. 188 pp. Extra Series XXXV. Cambridge: Cambridge University King, C.M. and Moody, J.E. (1982). The biology of the stoat Press. 1021 pp. (Mustela erminea) in the national parks of New Zealand. II. Bell, B.D. (1978). The Big South Cape Island rat irruption. Food habits. New Zealand Journal of Zoology 9: 57–80. Pp. 33–40. In: Dingwall, P.R., Atkinson, I.A.E. and Hay, Kuschel, G. (1971). Coleoptera: Curculionidae. Pp. 225– C. (eds). The ecology and control of rodents in New Zealand 260. In: Gressitt, J.L. (ed.). Entomology of the 34 Tuhinga, Number 29 (2018)

and other islands south of New Zealand. Pacific Insects ecology and control of rodents in New Zealand nature Monograph 27. Honolulu: Bishop Museum. 333 pp. reserves. Wellington: Department of Lands and Survey. Kuschel, G. (1982). Apionidae and Curculionidae 237 pp. (Coleoptera) from the Poor Knights Islands, New Ramsay, G.W. and Gardner, N.W. (1977). Endangered and Zealand. Journal of the Royal Society of New Zealand 12: rare New Zealand invertebrate species. The Weta 1: 3–6. 273–282. Reischek, A. (1888). Notes on rats. Transactions of the New Leschen, R.A.B., Marris, J.W.M., Emberson, R.M., Zealand Institute 20: 125–126. Nunn, J., Hitchmough, R.A. and Stringer, I.A.N. Russell, J.C., Towns, D.R., Anderson, S.H. and Clout, (2012). The conservation status of New Zealand M.N. (2005). Intercepting the first rat ashore. Nature Coleoptera. New Zealand Entomologist 35: 91–98. 437: 1107. McGuinness, C.A. (2001). The conservation requirements Taylor, R.H. (1978). Distribution and interactions of rodent of New Zealand’s nationally threatened invertebrates. species in New Zealand. Pp. 135–141. In: Dingwall, Threatened Species Occasional Publication 20. P.R., Atkinson, I.A.E. and Hay, C. (eds). The ecology Wellington: Department of Conservation. 658 pp. and control of rodents in New Zealand nature reserves. Martin, N.A. (2015). New Zealand flax –Phormium tenax Wellington: Department of Lands and Survey. 237 pp. (live plants). Landcare Research Plant-SyNZ, invertebrate Taylor, R.H. and Thomas, B.W. (1993). Rats eradicated identification chart, expert level (demonstration) 2015, from rugged Breaksea Island (170 ha), Fiordland, New p. 1. Retrieved on 30 August 2017 from http://plant- Zealand. Biological Conservation 65: 191–198. synz.landcareresearch.co.nz/Demonstartion_ID_Charts Thomas, B.W., Meads, M.J. and Notman, P.R. (1992). A /Phormium%20tenax%20(live)%20expert2015.pdf. report on the restoration of knobbled weevils (Hadramphus Meads, M.J. (1990). Forgotten fauna: the rare, endangered, and stilbocarpae) and flax weevils (Anagotus fairburni) to protected invertebrates of New Zealand. DSIR Information Breaksea Island, Breaksea Sound, Fiordland. Land Series 170. Wellington: DSIR. 95 pp. Resources Technical Record 79. Lower Hutt: Medway, D.G. (2011). The broad-billed prion (Pachyptila Department of Scientific and Industrial Research. 30pp. vittata) nesting colonies recorded during Cook’s visit to Thomson, G.M. (1921). Wild life in New Zealand. Part Dusky Sound in 1773. Notornis 58: 165–168. 1, Mammalia. Manual No. 2. Wellington: New Zealand Miskelly, C.M. (2014). Legal protection of New Zealand’s Board of Science and Art. 112 pp. indigenous terrestrial fauna – an historical review. Townsend, A.J., de Lange, P.J., Duffy, C.A.J., Miskelly, Tuhinga: Records of the Museum of New Zealand Te Papa C.M., Molloy, J. and Norton, D.A. (2008). New Tongarewa 25: 27–103. Zealand Threat Classification System manual. Miskelly, C.M., Tennyson, A.J.D., Edmonds, H.K. and Wellington: Department of Conservation. 35 pp. McMurtrie, P.G. (2017a). Dispersal of endemic Veale, A.J., Hannaford, O.D., Russell, J.C. and Clout, passerines to islands in Dusky Sound, Fiordland, M.N. (2012). Modelling the distribution of stoats on following translocations and predator control. Notornis New Zealand offshore islands. New Zealand Journal of 64: 192–205. Ecology 36: 38–47. Miskelly, C.M., Tennyson, A.J.D., Stahl, J.-C., Smart, A.F., Watt, J.C. (1975). The terrestrial insects. Pp. 507–535. Edmonds, H.K. and McMurtrie, P.G. (2017b). Breeding In: Kuschel, G. (ed.). Biogeography and ecology in New petrels of Dusky Sound, Fiordland – survivors from a Zealand. The Hague: W. Junk. 689 pp. century of stoat invasions. Notornis 64: 136–153. Wildland Consultants and Department of Conservation Peat, N. (2007). New Zealand’s fiord heritage: a guide to the (2016). Tamatea/Dusky Sound conservation and historic sites of coastal Fiordland. : Department restoration plan. N.p.: Department of Conservation. 155 pp. of Conservation. 76 pp. Peat, N. and Patrick, B. (1996). Wild Fiordland: discovering Unpublished sources the natural history of a world heritage area. Dunedin: University of Otago Press. 143 pp. Marris. J. (2001). Beetles of conservation interest from the Ramsay, G.W. (1978). A review of the effects of rodents Three Kings Islands. A report submitted to the on the New Zealand invertebrate fauna. Pp. 89–95. In: Department of Conservation, Northland Conservancy. Dingwall, P.R., Atkinson, I.A.E. and Hay, C. (eds). The Lincoln: Lincoln University. 28 pp. Tuhinga 29: 35–49 Copyright © Museum of New Zealand Te Papa Tongarewa (2018) 35

Alexander McKay: New Zealand’s first scientific photographer Simon Nathan

GNS Science, 1 Fairway Drive, Lower Hutt 5010, New Zealand ([email protected])

ABSTRACT: Alexander McKay (1841–1917) explored many parts of New Zealand while working as a fossil collector and geologist for the between 1873 and 1902. He was also a keen amateur photographer, taking photographs of geological features and documenting the impact of the 1888 and 1901 Canterbury earthquakes. He invented a telephoto lens about 1890, and later developed techniques for photographing fossils and microscopic thin sections of rocks. All of McKay’s varied photography was aimed at illustrating the scientific work he was undertaking, and as such he can be designated New Zealand’s earliest scientific photographer.

KEYWORDS: Alexander McKay, 1888 North Canterbury earthquake, 1901 Cheviot earthquake, scientific photography, photomicrographs, Awatere Valley, telephoto lens, historic photographs.

Introduction

Alexander McKay (1841–1917; Fig. 1) is a legendary figure among New Zealand geologists. As a young man, he emigrated from Scotland in 1863, hoping to make a living on New Zealand’s goldfields. James Hector, director of the New Zealand Geological Survey and Colonial Museum, employed him as a fossil collector in 1873 and soon recognised his scientific ability. Based at the Colonial Museum in Wellington, McKay became a pioneering geologist, gradually acquiring an encyclopaedic knowledge of New Zealand’s rocks, minerals and fossils (Bishop 2008). Largely self-taught, he relied on Scottish geologist Charles Lyell’s Principles of Geology (1830–33) and his own observations to develop geological interpretations. He was the first to recognise that New Zealand’s mountains were young and still rising, a concept not generally accepted until the mid-twentieth century. Reporting on the 1888 North Canterbury Fig. 1 Alexander McKay, aged about 50 (Cyclopedia earthquake, he was the first in the world to record and Company Ltd 1897: 174) (photo: McKee and Gamble, photograph transcurrent (sideways) movement on a fault Wellington). rupture. 36 Tuhinga, Number 29 (2018)

Fig. 2 Large cliff-face exposure of limestone at Weka Pass, North Canterbury. The controversial break between the Weka Pass Stone (above) and the Amuri Limestone (below), thought to represent the Cretaceous–Tertiary boundary, is indicated by the arrow. From microfossil studies, it is now known that the rocks above and below the contact are of early Oligocene age, and that the Cretaceous–Tertiary boundary is lower in the sequence. June 1886. GNS Science BW4882 (photo: Alexander McKay).

Throughout his career, McKay was interested in Cathedral Square, for which he paid £5 (Bishop 2008: 94). photography and took photographs to illustrate aspects of Difficult times meant that McKay had to undertake a his varied geological work. Although this has been noted by variety of labouring jobs over the next three years, and his previous writers (Cooper 1993; Bishop 2008), the full extent aspirations for a photographic career disappeared. In 1871, of McKay’s photographic work has never been documented. while prospecting a coal seam in the Ashley Gorge, he met In part, this is because the record is incomplete, and McKay geologist Julius Haast, who offered him a temporary job wrote only a single report on his photography (McKay as a field assistant. Based on Haast’s recommendation, in 1890b). I have assembled a list of the available information 1873 McKay was employed by Hector, initially as a fossil in the Appendix to this paper, and present a chronological collector and later as assistant geologist for the New Zealand account of his photography below. Geological Survey. He stayed in the same position, based at the Colonial Museum in Wellington but travelling Early photographic experience extensively around New Zealand, for the next 30 years. In 1867, McKay found employment at an isolated farm, There is no known record of any photographic work Lake Ohau Station in South Canterbury, where he met his undertaken by McKay through the 1870s, although future wife, Susannah. He was interested in setting up a draughtsman John Buchanan, who also worked for Hector business as a commercial photographer, and in early 1868 in the Colonial Museum, used photographs as a guide for he spent several weeks in Christchurch being tutored in his drawings of fossils and biological objects, and these may wet-plate photography at Edward Wheeler & Co. in have been taken by McKay. Alexander McKay: New Zealand’s first scientific photographer 37

During the early 1880s, the wet-plate process was these may have been taken as early as 1884, but we cannot superseded by the development of mass-produced dry be sure of any dates apart from those of the Weka Pass plates, which meant that it was feasible to take photographs group, which were taken when McKay visited the area on of geological features in the field and later process the plates 14–16 June 1886 (McKay 1887). in a darkroom. Hector wrote to his London agent, James During the late nineteenth century, there were ongoing Farmer, asking him to procure a portable camera suitable arguments among New Zealand geologists about the for geological fieldwork. Farmer responded to Hector on location of the Cretaceous–Tertiary boundary (Oldroyd 13 June 1883, saying that he had obtained ‘a complete set of 1972). Although this is clearly marked by an unconformity Photographic Apparatus’, and dispatched it to New Zealand in classic sections in the United States, Britain and Europe, by sea (Mildenhall & Nathan 2013: 78; Nathan 2015: there appeared to be no break in New Zealand. Hector 151–152). It is tantalising that we have no details of the believed that the contact occurred at a minor break equipment, but it seems likely that it included the camera between the Amuri Limestone and Weka Pass Stone in McKay used for his photography through the late 1880s. North Canterbury, but this was contested by Haast and Items 1 and 2 in the Appendix list approximately 200 Canterbury College geologist Frederick Hutton. Hector photographs of geological features taken by McKay before sent McKay to investigate the well-exposed sequence of 1892. The majority were taken in 1888–89 as part of beds at Weka Pass, and McKay took a series of photographs investigations into the 1888 North Canterbury earthquake to supplement his report (McKay 1887), one of which is (see below), but there are also small groups of photographs reproduced as Fig. 2. Unfortunately, it was too expensive to from Northland, Wellington, Nelson, the Mokihinui publish photographs at the time, and the report is illustrated coalfield and the Weka Pass area in Canterbury. Some of by sketches and cross sections. 38 Tuhinga, Number 29 (2018)

1888 North Canterbury earthquake over the summer of 1888/89 to join up with the work he had done earlier in the Clarence Valley (McKay 1886). A large earthquake in North Canterbury on 1 September Hector was clearly convinced by the value of photography, 1888 caused considerable damage to the area around because McKay returned to Marlborough with his camera, Hanmer. McKay was sent to investigate, and left Wellington a supply of glass plate negatives and instructions about some on 27 September. To reach the damaged area, he travelled features to be photographed. From November 1888 to up the Awatere Valley, where he was able to examine the February 1889, McKay took more than 100 photographs fault rupture on the Awatere Fault caused by the 1848 of geological features and the 1848 rupture of the Awatere Marlborough earthquake. Grapes (2009, 2011: chapter 11) Fault. He also returned to the middle Clarence Valley to gives a detailed day-by-day account of the journey and photograph geological features he had seen during earlier McKay’s impressions. The line of the 1848 rupture was still fieldwork (Hector 1890: xxxvi). It was probably towards clear, and McKay noted that it had not broken again in the end of this period that McKay travelled south to the 1888. As he travelled south over Jollies Pass, the damage area affected by the 1888 earthquake around the Waiau increased, until he found widespread ground rupture near River, photographing the damage and rupturing along the the Waiau River along the line of what we now know as the Hope Fault that he had observed several months earlier. A Hope Fault. At Glynn Wye Station, McKay noted sideways selection of these photographs is shown in Figs 3–9: Figs (transcurrent) offset of 2–3 m, marked by offset fences. He 3 and 4 show the 1888 rupture on the Hope Fault; Fig. 5 wrote up his report promptly on his return to Wellington shows the 1848 rupture on the Awatere Fault; and Figs 6–9 in October 1888. It appears that he did not have a camera show different aspects of the geology. with him during this trip, because he comments that ‘No In referring to McKay’s report on the 1888 earthquake, pen, and scarce the most graphic touches of a graphic artist Hector (1890: xxxi) noted that ‘Mr McKay’s report gives could give a just and complete rendering of the river bank the results of his observations made on the spot, and he has … Photography properly directed might do something’ supported them by means of a large series of well-selected (McKay 1890a: 6). photographs, which will be published as soon as possible’. McKay was back in Marlborough in November 1888. Unfortunately, this did not happen, and it more than a Hector had decided that he should create a geological map decade passed before a small group of photographs was the Awatere Valley and the adjacent part of east Marlborough published (McKay 1902).

Fig. 4 Figure lying on the ground indicating the size of the vertical offset on the 1888 fault rupture (about a metre) – considerably less than the 2.6 m horizontal offset. GNS Science BW4770 (photo: Alexander McKay). Alexander McKay: New Zealand’s first scientific photographer 39

Fig. 5 Prominent scarp and trench between Lee Brook and Castle River in the upper Awatere Valley. This marks the line of the rupture along the Awatere Fault during the 1848 earthquake. When it was photographed 40 years later, there was little erosional modification. GNS Science BW4945 (photo: Alexander McKay).

Fig. 6 Entrance to the limestone gorge in Swale Stream, Clarence Valley. Late Cretaceous limestone dips (slopes) away from the viewer at about 45 degrees. A, view in 1888–89; GNS Science BW4756 (photo: Alexander McKay). B, view in 2008 (photo: James Crampton). Since the earlier photograph was taken, aggradation has built up the level of the stream, and there have been further changes following the 2016 Kaikoura earthquake. 40 Tuhinga, Number 29 (2018)

Fig. 7 Exposure of steeply dipping Great Marlborough Conglomerate (Miocene) in Hevers Creek, near Kekerengu, containing large blocks of Amuri Limestone. The widespread occurrence of this unit persuaded McKay that there had been uplift of mountains and erosion over the last 10 million years. GNS Science BW 4885 (photo: Alexander McKay).

Fig. 8 Steeply dipping basalt dikes cutting across gently dipping Cretaceous sediments, Winterton River (a tributary of the Awatere River). GNS Science 4885 (photo: Alexander McKay). Alexander McKay: New Zealand’s first scientific photographer 41

Fig. 9 Large inoceramid bivalve fossil in Cretaceous sediments. Location uncertain. GNS Science BW4971 (photo: Alexander McKay).

Development of a telephoto lens the glass screen at the back of the camera, as well as the problems of eliminating fuzziness caused by spherical and On 13 August 1890, McKay gave an address to the chromatic aberrations. Patterson (2009) gave a modern Wellington Philosophical Society titled ‘On some means summary of how McKay had prepared an extension tube of increasing the scale of photographic lenses and the that could be attached to the camera, but that was still use of telescopic powers in connection with an ordinary portable enough to be taken into the field. camera’ (New Zealand Institute 1890). In essence, this One of the best-known examples of a photograph taken was a description of how he had developed a telephoto by McKay with his telephoto lens is of a ship in Wellington lens that could be attached to his camera (Fig. 10). McKay Harbour from Tinakori Road near the Shepherds’ Arms illustrated the talk with examples of photographs he had Hotel (Fig. 11). The vessel was previously believed to be taken. There was great interest from the audience, and the Russian warship Vestnik, which visited Wellington Hector commented that it was the product of years of work in 1886 (Main 1972: 101; Bishop 2008: 13, 157–158), by McKay and considerable expense. McKay subsequently but this appears to be based on an incorrect inference. A published the talk as a paper in the Transactions of the New note about the image in McKay’s handwriting, in item 2 Zealand Institute, but unfortunately without any illustrations in the Appendix, refers to the ‘Flag ship of the German (McKay 1890b). The difficulties of magnifying distant Squadron’; this has been identified by Perrin (2017) as SMS objects had long been a matter of frustration to McKay, Sophie, which visited Wellington in November–December and his solution was to design the lens like a telescope. 1890. The date is important, as it indicates when McKay He described the technical issues of getting enough light was getting good results with the telephoto lens. It does through the lens so that the image could be focused on not appear that McKay was using his telephoto lens much, 42 Tuhinga, Number 29 (2018)

if at all, before 1890. The large batch of photographs he took in North Canterbury and Marlborough in 1888–89 does not include any telephoto shots. There is little doubt that McKay developed his telephoto lens independently, and in his entry in the Cyclopedia Company Ltd (1897: 175) he claimed to be its inventor. Patterson (2009: 365) summarises attempts to develop a practical telephoto lens in Britain and Europe, culminating in a patent application by Thomas Dallmeyer of London in 1891. The issue of priority is open to debate, but McKay was clearly a pioneer in this area of photographic technology. Fig. 10 Telephoto lens developed by Alexander McKay, and described in his 1890 paper. Private collection (photo: Photography through Hamish Campbell). a microscope In 1892, McKay was given increased responsibility. As part of local geologists about the nomenclature of the fine-grained a reorganisation under the new Minister of Mines, Richard rocks, which were often hydrothermally altered (McKay Seddon, the control of geological work was transferred from 1900). McKay decided that the only way to resolve the Hector to the Undersecretary of Mines (Nathan 2015: 196), issue was to submit his rock collection to an overseas who delegated this work to McKay. His title was initially expert for examination by microscope. William Sollas, Mining Geologist, later Government Geologist. For the next FRS, a professor of geology at the University of Oxford, decade, McKay worked almost entirely on economic geology. was contracted by the New Zealand Mines Department to The Coromandel Peninsula (then also known as the Cape provide a detailed sample-by-sample report on 500 rocks Colville Peninsula) had become the centre of underground submitted to him by McKay – 402 from the Coromandel gold mining in New Zealand. In 1896–97, McKay undertook a Peninsula, and 98 from other parts of New Zealand. Rock major survey of the region, discussing the geology and mineral slices (thin sections) were prepared from the samples in prospects, making a large rock collection (McKay 1897) and Britain for petrological examination by Sollas using a taking photographs. Almost all the rocks in the region are polarising microscope. The rocks and thin sections were volcanic, and there was considerable disagreement among subsequently returned to New Zealand.

Fig. 11 (below) Photographs taken from the same position above Tinakori Road, Wellington, looking towards the harbour on the same day in November–December 1890. A, view with normal lens, with Shepherds’ Arms Hotel in the foreground. The masts of a ship can be seen faintly in the upper centre. Alexander Turnbull Library ½-021850-F (photo: Alexander McKay). B, view with telephoto lens showing the SMS Sophie, part of the visiting German squadron, about 2.5 km away. Alexander Turnbull Library ½-021856-F (photo: Alexander McKay). Alexander McKay: New Zealand’s first scientific photographer 43

Fig. 12 Examples of photomicrographs published as full-page illustrations in The rocks of the Cape Colville Peninsula. Vol. 1 (Sollas & McKay 1905). A, hornblende dacite from Cabbage Bay, with large, clear phenocrysts of plagioclase and a single large phenocryst of dark brown hornblende (centre) (photo: Alexander McKay). B, volcanic conglomerate band in pre-Cenozoic greywacke, from a road cutting near Coromandel. The thin section shows that the pebbles include a variety of volcanic lithologies (photos: Alexander McKay).

While there is no evidence that McKay undertook any Colville Peninsula have received high commendation from microscope petrography himself, he turned his attention to the Director of the Geological Survey in Great Britain, photographing thin sections. When Sollas’s reports were who in a letter of acknowledgement expressed a doubt received, McKay wanted to ensure that they were printed whether such a class of work could be done in the Old in full, with adequate illustrations. The resulting two- Country’ (Anon. 1906). volume monograph, The rocks of the Cape Colville Peninsula One of the unusual rocks described by Sollas was a (Sollas & McKay 1905, 1906) is one of the largest and most peculiar streaky rhyolitic rock that was found around elaborate publications on New Zealand geology, containing Waihi, and was known locally as wilsonite. Sollas was 32 landscape views and 177 full-page photographs of thin clearly puzzled by (Sollas & McKay 1905: 124– sections (photomicrographs), all taken by McKay (examples 125), which we now recognise as a variety of ignimbrite are shown in Figs 12 and 13). It was recorded in the 1906 New containing welded and flattened pumice fragments. Zealand Mines Record that his ‘greatly enlarged photographs of McKay’s photographs (Fig. 13) are possibly the first rock slides accompanying his work on the The rocks of the Cape published photographs of this rock type. 44 Tuhinga, Number 29 (2018)

Fig. 13 Wilsonite, now recognised as a variety of ignimbrite, published in The rocks of the Cape Colville Peninsula. Vol. 1 (Sollas & McKay 1905): A, cliffs of Wilsonite in a quarry in the Ohinemuri Valley, near Waikino; B, cut slab showing flattened pumice fragments in a fine-grained matrix; C, thin-section view, again showing flattened pumices and a few phenocrysts. (photos: Alexander McKay). Alexander McKay: New Zealand’s first scientific photographer 45

Fig. 14 Two examples of earthquake damage during the 1901 Cheviot earthquake, published in Alexander Mackay’s account of the earthquake (McKay 1902): A, the rear of Penberthy’s Hotel, Cheviot, after the earthquake of 16 November 1901; B, damage to the bridge over Swamp Creek, Sinclair Road, Cheviot (photos: Alexander McKay).

1901 Cheviot earthquake single-storey wooden bungalows, and the damage was not overwhelming, although most chimneys had collapsed. On 16 November 1901, the North Canterbury region McKay had his camera with him and took 76 was shaken again by a large earthquake that was centred photographs of the damage around Cheviot (Bishop 2008: near the town of Cheviot, with aftershocks continuing 173). His report on the earthquake was written up promptly for more than two months. McKay was sent to investigate for publication by the Government Printer (McKay 1902) and report on the impact of the earthquakes. Based on his and included 12 photographs of damage around Cheviot experience in Marlborough, he looked for a major fault in 1901 (Fig. 14). McKay took the opportunity to include rupture, but was unable to find a geological cause for the parts of the text of his report on the 1888 earthquake, as earthquakes. The buildings in Cheviot were predominantly well as six previously unpublished photographs. 46 Tuhinga, Number 29 (2018)

Fig. 15 Photographs of fossils held in the John Buchanan collection, GNS Science. It is likely that the photographs were taken to assist Buchanan in making drawings of fossils and biological specimens. A, fossil gastropod, partly enclosed in rock matrix. It is possible that McKay coated the fossil with fumes of ammonium chloride, giving a white coating that reduces shadows (photo: Alexander McKay). B, different views of a fossil oyster specimen. (photo: Alexander McKay).

Photographing collections as his most significant contribution to New Zealand geology (Bishop 2008: 179), and had experimented fossils with photographing fossils (Fig. 15). Hector had intended In March 1902, McKay was feeling unwell. His doctor to have the fossils identified and described, but this never diagnosed heart problems, and stated that he was no longer happened. In 1904, after Hector’s retirement, McKay fit enough to undertake fieldwork. This marked the started the enormous task of sorting out and repacking the beginning of the final stage of McKay’s career, when he fossils he had collected so that they could be moved out of was confined to the office. the museum. It must have been frustrating for McKay to McKay had originally been employed by Hector as see the fruits of his collecting being moved into storage, a fossil collector, and over the years had amassed a huge and he experimented with photographing the fossils so that volume of fossil material that was stored within and under an accessible record would be available. In a memorandum the Colonial Museum. McKay himself regarded his fossil to the Undersecretary of Mines, he mentions - Alexander McKay: New Zealand’s first scientific photographer 47

a style of photography which has not previously been Conclusion accomplished. Hitherto the photographing of natural objects, such as rocks and fossils, has, in comparison Although Alexander McKay is renowned as a geologist, with drawings laboured under the disability of being his name is little known among nineteenth-century New Zealand photographers. Both Knight (1971) and Main accompanied by shadows that, seemingly unavoidable (1972) note his development of the telephoto lens, but in connection with the photograph, are not necessarily there is no mention of his other photography. a part of the drawing. To take photographs that would In McKay’s busy life, photography was largely an not show the ordinarily accompanying shadow became activity he carried out in his spare time, but because of his with me first a dream, next a matter of experiment, and keen interest he achieved high-quality results. There is no finally an accomplished fact. I know not whether I evidence that he took any personal or group images, and it should claim for my achievement the rank of a appears that he regarded photography essentially as a way of discovery, but certain it is that it enables the production recording the geological features he observed as well as the of process blocks equal to those from drawings, and at impact of events such as the 1888 and 1901 earthquakes. Over the years, McKay extended his techniques to much less cost (Anon. 1906). include telephoto and scientific photography. He used his Unfortunately, McKay gave no technical details of how he photography to illustrate scientific features, and can thus be obtained his shadowless photographs. His original memo categorised as a scientific photographer – probably the first was accompanied by 37 photographs of rocks and minerals, who can hold that title in New Zealand. The ethnologist but these are not included in the file held in Archives New Augustus Hamilton was also taking scientific photographs Zealand in Wellington (R17869560). in the 1890s (McCredie 2017), but McKay appears to have started several years earlier. When McKay started taking photographs in the 1880s, What happened to there was little possibility of publishing the images – it was simply too expensive, and line drawings on woodcuts were McKay’s photographs? used instead. Although McKay’s photographs of the 1888 earthquake would have been of immediate interest, James It is tantalising that there is so little technical information Hector had no funds available for their publication, either about how McKay took his photographs or the equipment in the Reports of Geological Exploration or the Transactions he used. Although money for photographic material was of the New Zealand Institute. A decade later, when McKay presumably supplied by the New Zealand Geological was Government Geologist, and he took the opportunity Survey or Mines Department for his scientific photographs, to ensure that his report on the 1901 Cheviot earthquake he apparently retained all the negatives and most of the prints. was well illustrated, including with it his photographs of In McKay’s will, dated 10 October 1916 – some months the 1888 earthquake, which had never been published. The before his death – he left ‘all my photographic apparatus, rocks of the Cape Colville Peninsula (Sollas & McKay 1905, material for producing photographic prints, and all my 1906) is a landmark publication because it is one of the photographic negatives’ to his older son, William McKay earliest scientific monographs illustrated by photographs, (McKay n.d.). The photographic prints were to be split entirely produced in New Zealand by the Government between his widow, Adelaide (McKay had remarried Printer. In addition to his talent as a scientific photographer, following the death of Susannah), and his younger son, McKay had the drive and organising ability to ensure that Duncan. A few items, including the telephoto lens, have his illustrated manuscripts were completed and published. been passed down the family, but almost none of the photographs have been located. The evidence (as detailed in the Appendix) indicates that McKay took more than 500 photographs through his career, and it is hoped that publication of this paper may assist in the identification of more of his photographs if they have survived. 48 Tuhinga, Number 29 (2018)

Acknowledgements McKay, A. (1887). On the junction of the Amuri Limestone and Weka Pass Stone, north Canterbury. Reports of I am grateful to Philip Carthew for his assistance in Geological Exploration 18[1886-87]: 78–91. preparing the photographs for publication and to Rowan McKay, A. (1890a). Preliminary report on the earthquakes Burns for a preliminary catalogue of the Alexander McKay of September 1888 in the Amuri and Marlborough districts of the South Island. Reports of Geological photographs held at GNS Science. Wayne Barrar, Hamish Exploration 20[1888–89]: 1–16. Campbell, Rodney Grapes, Athol McCredie and Nick McKay, A. (1890b). On some means for increasing the scale Perrin are thanked for useful discussion on different aspects of photographic lenses and the use of telescopic powers of the photographs. The Real McKay (2008) by the late in connection with an ordinary camera. Transactions of Graham Bishop was an invaluable source of biographical the New Zealand Institute 23: 461–465. information about McKay. McKay, A. (1897). Report on the geology of the Cape Colville Peninsula, Auckland. Appendices to the Journal of the House of Representatives C-9: 1–75. McKay, A. (1900). Importance of the correct naming of rocks of the Cape Colville Peninsula. New Zealand Mines Record III(8): 305–309. References McKay, A. (1902). Report on the recent seismic disturbances within Cheviot County in northern Canterbury and in Anon. (1906). [Untitled article]. New Zealand Mines Record the Amuri district of Nelson, New Zealand (November and 14(8): 360. December 1901). Wellington: Government Printer. 80 pp. Bishop, D.G. (2008). The real McKay: the remarkable life Mildenhall, E. and Nathan, S. (2013). Letters from James of Alexander McKay, geologist. Dunedin: Otago Farmer to James Hector, 1876–1891. Geoscience University Press. 252 pp. Society of New Zealand Miscellaneous Publication Cooper, R. (1993). McKay, Alexander. Dictionary of New 133I. Wellington: Geoscience Society of New Zealand. Zealand biography [website]. Retrieved on 1 March 133 pp. 2018 from https://teara.govt.nz/en/biographies/2m12/ Nathan, S. (2015). James Hector: explorer, scientist, leader. Lower Hutt: Geoscience Society of New Zealand. mckay-alexander. 264 pp. Cyclopedia Company Ltd (1897). Cyclopedia of New New Zealand Institute (1890). Report on meeting of the Zealand. Vol. 1, Wellington Provincial District. Wellington Philosophical Society on 13 August 1890. Wellington: Cyclopedia Company Ltd. 1373 pp. Transactions and Proceedings of the New Zealand Institute Grapes, R. (2009). Alexander McKay and the Awatere 23: 605–606. Fault, New Zealand: ground rupturing and large scale Oldroyd, D.R. (1972). Nineteenth century controversies horizontal displacement. New Zealand Journal of concerning the Mesozoic/Tertiary boundary in New Geology and Geophysics 52: 349–364. Zealand. Annals of Science 29(1): 39–57. Grapes, R. (2011). The visitation: the earthquakes of 1848 Park, J. (1906). The geology of the area covered by the and the destruction of Wellington. Wellington: Victoria Alexandra sheet, Central Otago Division. New Zealand University Press. 191 pp. Geological Survey Bulletin 2. Wellington: New Zealand Hector, J. (1890). Progress report, 1888–89. Reports of Geological Survey. 52 pp + 8 maps. Geological Exploration 20: ix–lvii. Patterson, J. (2009). Alexander McKay and development of Knight, H. (1971). Photography in New Zealand: a social the telephoto lens. New Zealand Journal of Geology and and technical history. Dunedin: J. McIndoe. 196 pp. Geophysics 52: 364–365. Main, W. (1972). Wellington through a Victorian lens. Perrin, N. (2017). Alexander McKay’s first telephoto. Wellington: Millwood Press. 108 pp. GSNZ Journal of the Historical Studies Group 57: 23–27. McCredie, A. (2017). Augustus Hamilton – creating a visual Sollas, W.J. and McKay, A. (1905). The rocks of Cape database. Journal of the Royal Society of New Zealand Colville Peninsula, Auckland, New Zealand. Vol. 1. 47(1): 138–144. Wellington: Government Printer. 289 pp. McKay, A. (1886). On the geology of the eastern part of Sollas, W.J. and McKay, A. (1906); The rocks of Cape the Marlborough provincial district. Reports of Geological Colville Peninsula, Auckland, New Zealand. Vol. 2. Exploration, 16[1885]: 27–136. Wellington: Government Printer. 215 pp. Alexander McKay: New Zealand’s first scientific photographer 49

Unpublished sources Society on 8 August 1891. The brief descriptions include the images taken in 1888–89 in Marlborough noted above, McKay, A. (n.d.). Probate file for Alexander McKay, but also small groups of photographs from Northland, the geologist, Wellington. R23027928, Archives New Mokihinui coalfield, Nelson and Wellington. None of Zealand, Wellington. the lantern slides has been located.

Appendix: 3 Alexander Turnbull Library, Wellington Photographs of six prints (021849, 021850, 021851, Inventory of photographs taken 021855, 021856 and 021857), copied from an album loaned by Mrs A.H. Parker, Hukanui, in 1963. These by Alexander McKay include four photographs taken by McKay in Tinakori Listed here are the various sources from which I have been Road, Wellington, demonstrating his telephoto lens, as able to discover information on photographs taken by well as images of a settlement in the Manawatu Gorge Alexander McKay. None of his original glass plate negatives and the hotel at Mitchells, West Coast. There is also a has been located, and the photographs that survive are either single image (051913) showing the Kelburn School under prints, copies of prints or images in publications. construction. 4 Report on the recent seismic disturbances within 1 GNS Science, Lower Hutt Cheviot County (McKay 1902) GNS Science holds a collection of 147 half-plate contact In the first part of this document, McKay repeats part prints, mounted on card, that were clearly intended of his report on the 1888 earthquake (McKay 1890a), for display. Many of the prints have labels on the back, and includes six images that had not been previously some in McKay’s handwriting. There are a number of published. Prints of these are held at GNS Science. He duplicates – the older ones are albumen prints, with a records having taken 76 photographs of the area affected sepia tint, and later ones are on silver gelatin printout by the 1901 earthquake (Bishop 2008: 173), of which 12 paper (Wayne Barrar, pers. comm., 2017). The majority are reproduced in the report. None of the originals of this of the prints appear to have been taken in 1888–89, large group has been located. and feature the Clarence and Awatere valleys, as well as 5 The rocks of the Cape Colville Peninsula earthquake damage in the Waiau Valley and especially (Sollas & McKay 1905, 1906) around Glynn Wye Station. There are also small groups of photographs taken at Weka Pass and around Red All the photographic illustrations in these two volumes Rocks in Wellington. The original mounted prints are were apparently taken by McKay. There are 32 ground held in the security room at the GNS Science Library. shots of different parts of the region in Volume 1, as well They have been photographed, and the negatives are in as a total of 218 photomicrographs of thin sections of the GNS Science photo collection, listed as 4616–4627, different rocks in both volumes. 4751–4762, 4766–4795, 4844–4867, 4875–4902 and 4936–4976. A small collection of prints of photographs of 6 New Zealand Geological Survey Bulletin 2 fossils is also held in the John Buchanan collection; some (Park 1906) of these were subsequently used as the basis for drawings This volume contains 17 photomicrographs of thin by him. sections from the region around Alexandra in Otago (facing pages 46 and 48) taken by McKay. 2 Hocken Collections, University of Otago,

Dunedin The Hector papers contain a memo in McKay’s handwriting (Ms-0445-2/13) titled ‘Lantern Slides of New Zealand Scenery by Alexander McKay’, consisting of a list of 173 photographs. It seems probable that McKay’s negatives were converted to lantern slides for the talk that Hector and McKay gave to the Wellington Philosophical Tuhinga 29: 50–76 Copyright © Museum of New Zealand Te Papa Tongarewa (2018)

Rising sun at Te Papa: the Heriot collection of Japanese art David Bell* and Mark Stocker**

* University of Otago College of Education, PO Box 56, Dunedin, New Zealand ([email protected])

** Museum of New Zealand Te Papa Tongarewa, PO Box 467, Wellington, New Zealand ([email protected])

ABSTRACT: In 2016, the Museum of New Zealand Te Papa Tongarewa (Te Papa) acquired more than 60 Japanese artworks from the private collection of Ian and Mary Heriot. The works make a significant addition to the graphic art collections of the national museum. Their variety and quality offer a representative overview of the art of the Japanese woodblock print, and potentially illustrate the impact of Japanese arts on those of New Zealand in appropriately conceived curatorial projects. Additionally, they inform fresh perspectives on New Zealand collecting interests during the last 40 years. After a discussion of the history and motivations behind the collection, this article introduces a representative selection of these works, arranged according to the conventional subject categories popular with nineteenth- and early twentieth-century Japanese audiences. Bijin-ga pictures of beautiful women, genre scenes and kabuki-e popular theatre prints reflect the hedonistic re-creations of ukiyo ‘floating world’ sensibilities in the crowded streets of Edo. Kachöga ‘bird and flower pictures’ and fukeiga landscapes convey Japanese sensitivities to the natural world. Exquisitely printed surimono limited editions demonstrate literati tastes for refined poetic elegance, and shin-hanga ‘new prints’ reflect changes in sensibility through ’s great period of modernisation. In sum, these works offer holistic appreciations of the diversity of pictorial interests, the technical and aesthetic triumphs of the polychrome woodblock print, and the emergence and ability of Japanese arts to engage international viewers today.

KEYWORDS: art collecting, Edo, Japanese art, ukiyo-e, nishiki-e, shin-hanga, bijin-ga, kabuki-e, fukeiga, kachöga, surimono.

Introduction

In 2016, the Museum of New Zealand Te Papa Tongarewa (Te Papa) purchased an art collection predominantly of Japanese prints – just over 60 in number – but also including sumi-e (ink-wash) paintings and books, from Ian and Mary Heriot, a retired couple living in Katikati (Fig. 1). Known informally as the Heriot collection, its inspiration, general scope and selected highlights will be the main subject of this article.

Fig. 1 Mary and Ian Heriot, Katikati, 2017 (Ian Heriot). Rising sun at Te Papa: the Heriot collection of Japanese art 51

The need to expand and invigorate Te Papa’s holdings early twentieth-century picturesque meishö-e landscape and in the area up to that point seems indisputable, particularly genre scenes. The collection embraces almost every format in hindsight. The older collection could be appropriately of print, from the printed book, exercises in drawing and described as small but interesting, comprising just 18 ukiyo-e, learning, hashira-e pillar prints and kakemono-e vertical scroll ‘floating world pictures’ from the Edo period (1603–1868). prints, the predominant öban single-sheet format (although The overwhelming majority of these prints consist of small- sometimes joined and forming diptychs or triptychs), to the scale meishö-e (pictures of famous places) by Andö Hiroshige highly refined art of the surimono, ‘rubbed thing’ limited- (1797–1858), together with a smattering of depictions of the edition poetic print, an area of particular fondness for Ian Yoshiwara licensed pleasure quarters.1 Two further prints Heriot.2 Within this rich field, we can trace the technical by Katsushika Hokusai (1760–1849) are almost certainly developments of woodblock printing, the emergence and later productions by the house of Watanabe Shözaburöō sustenance of a school system in the Katsukawa and Utagawa (1885–1962), from the early twentieth century. While this schools, and the development of stylistic characteristics of body of work provided potentially useful material to inform decorative elegance, complemented by technical virtuosity. exhibitions representing New Zealand collecting interests The Heriot collection not only provides rich insights into in the first half of the century, and perhaps even to inform more than two centuries of the craft, history and aesthetic trends in early modern stylistic developments of this country’s sensibilities of Japanese popular arts in all their variety, but artists, in itself it was of insufficient critical mass and quality it also reflects the shifting tides of Japanese and Western to comprise a discrete, coherent exhibition. Moreover, as intercourse before, during and beyond the turning point a representation of ukiyo-e in a national museum, it was of the Restoration (1867–68). The international manifestly inadequate. significance of Japanese arts during this period is evident in the patterns of collecting in New Zealand, which A catholic collection were consistent with those in western Europe and North America. The Heriots form part of a tradition of collecting A desire to inject into Te Papa’s collection a better sense Japanese print matter in New Zealand established early on of the greatest period of polychrome craft printmaking by Sir George Grey (1812–98), continued into the early in history helped inspire the recent acquisition. But there twentieth century by Sir Joseph Kinsey (1852–1936) and are other remarkable aspects to the Heriot collection itself more recently – up to the 1970s – through the development that also commended it. First and foremost, the scale of the of extensive private collections, especially in Canterbury, by collection in its entirety made it possibly the most extensive Francis Shurrock (1887–1977), Keith Mosheim (1919–87), one of its kind in private hands in New Zealand in the Gordon McArthur (1925–2010), Ronald J. Scarlett (1911– twenty-first century. Quantitatively, above and beyond Te 2002), and Greggory Kane (1921–78), many of whose items Papa’s selections, it surpasses the public collections in the were subsequently dispersed to public institutions.3 Auckland Libraries, Auckland Art Gallery Toi o Tämaki and

MTG Hawke’s Bay. In turn, it interestingly complements Ian Heriot as a collector Canterbury Museum’s impressively catholic holdings and the great masterpieces (notably Hokusai’s Thirty-six Views of Before examining specific areas of the collection, it is Fuji) in the Dunedin Public Art Gallery, with its strength in important to provide personal insights into the Heriots later print movements. themselves, and their ‘ruling passion’ as collectors. Te Papa Indeed, the Heriot collection represents the temporal aims to bring out ‘the people, places and ideas that have scope of Japanese print histories, from early monochrome shaped New Zealand’,4 and here they do not disappoint. illustrative matter, through to the crystallisation of Although Mary Heriot has herself in recent years become polychrome nishiki-e ‘brocade pictures’ by Suzuki Harunobu an enthusiastic collector of Japanese ceramics and has, on (1724–70) and Isoda Koryüsai (1735–90), the emergence occasion, good-naturedly reined in her husband’s ambitions of subject specialisations in bijin-ga ‘pictures of beautiful to make still further purchases, Ian Heriot has largely driven women’, kacho-e ‘bird and flower’ compositions, kabuki-e and defined the collection. He bought his first Japanese theatre pictures and actor representations, and Meiji era prints from a Takapuna second-hand shop in 1979, which shin-hanga ‘new picture’ print revivals, and extending to included Actor Posing with a Sword (c. 1835) by Utagawa 52 Tuhinga, Number 29 (2018)

Kunisada (Utagawa Toyokuni III) (1786–1865), subsequently “For what they are, they are quite decorative and nice, but purchased by Te Papa.5 A professional firefighter, then in his I wouldn’t take one of them home.” Because they were early 40s, and a father of three from a Scottish working- just a trend, these were churned out … most of those pretty class background, Ian had no grounding in art other than a girl things were churned out as ads.’ In contrast, Ian looks modest commercial interest in antique European furniture. out for ‘an unusual subject, or a subject that’s portrayed in Neither, till then, did he have any knowledge of ukiyo-e and a particular way. I never set out to buy the average, pretty admits, ‘I was naive enough to think I had stumbled on a Japanese print.’9 new form of art when I saw these prints.’6 His funds were More often than not, it was the scarcity of high-quality limited, but his enthusiasm was not. or interesting prints that frustrated Ian, who operated on a Although he would never classify himself ‘as any sort limited budget and, for the first 20 years, in a pre-Internet of an intellectual’, Ian’s curiosity was immediately fired by world. Opportunities to buy from the catalogues of offshore the prints and his interest in the area steadily deepened. dealers such as Hotei in the Netherlands were frustratingly Becoming seriously acquainted with the field ‘took a bit limited, as by the time these reached New Zealand, the of study. I got a print, I took it home and I worked on desired print – in many cases unaffordable anyway – was the print … and it gradually came to mind that I couldn’t already sold. Ian made a virtue of necessity by acquiring find another example.’ Establishing its rarity, perhaps even the overwhelming majority of his collection nationally, and uniqueness, was ‘quietly enjoyable’ for Ian. He found it this made genuine finds all the more exciting. Japanese arts surprising that many others did not operate the same way: had enjoyed great popularity with New Zealand collectors ‘It amazes me when I’ve struck some collectors of Japanese during the early twentieth century, and went through a art [who] don’t feel the need to learn anything about it.’7 resurgence in the late 1940s, but their market presence had Heriot was necessarily an autodidact – there was never waned by the 1980s. One highlight was when Ian, driving a critical mass of scholars or collectors in New Zealand, through Parnell, screeched to a stop when he noticed even in Auckland, where he spent his nearly 30 years of his a humble pavement vendor who had original-edition working life – who shared his interest. Through the years, Harunobu and Koryüsai prints for sale. This was a rare he has built up a substantial library of more than 500 books moment when he sacrificed his normal requirement for a on Japanese history, art and travel. His immersion in things print in at least fair condition in favour of the famous name, Japanese – including an ability to work out titles, names knowing the extreme scarcity of such works worldwide. and other relevant words on the prints – was accompanied Ian claims, with no false modesty, ‘Considering the funding by an increasingly discerning eye towards expansion of the basis behind me, which was virtually nil, and a certain collection. A reflection of the respect in which he is now amount of thoughtfulness going into the collecting of these held in the field is his and Mary’s role as long-standing prints, I think I’ve put together a collection that won’t moderators for the discussion board of Gotheborg.com, readily be done again.’10 which advises collectors on Japanese and Chinese antique porcelain.8 Leading dealers in prints, such as Sandra Andacht Collector’s choices and Robyn Buntin, are well known to the Heriots and vice versa. Ian Heriot asserts, ‘I’d like to think that I’m a reasonably Perhaps Ian was able to use his isolation to his own careful buyer of prints. They’ve got to have an aesthetic that advantage when dealers – probably best left unnamed – ‘did I immediately appreciate.’11 This matters more to him than not have a clue’ about what was hanging on their walls. genre, period, style or format. It is therefore not surprising As he explains, ‘Most dealers in New Zealand can look at that the magisterial and essentially canonical approaches something and say, “That’s a Japanese print”, and they may maintained through the generations by authorities such as be able to identify a Kuniyoshi or a Hokusai … but if it’s off Arthur Davison Ficke (1883–1945), Richard Lane (1926– the beaten track, the slightest bit, they’re lost.’ Sometimes a 2002) and, to a lesser extent, Frank Whitford (1941–2014) dealer can have a glut of uninteresting prints: ‘Peter Webb have little traction for Ian.12 From an early stage he was rang me up and said, “Would you give me some idea of attracted to considerably later Meiji era prints and early to what these prints are?” So, I went in and there were around mid-twentieth-century shin-hanga prints, especially those by about 150, all individually framed, and they were all of Hasui Kawase (1883–1957). Indeed, given the choice again, beautiful women [bijin-ga]. I went through the whole lot he says, ‘I think I’d replace a large part of my collection with and he said to me, “What do you think of them?” I said, early shin-hanga if I could afford them.’ Yet he is essentially Rising sun at Te Papa: the Heriot collection of Japanese art 53

pragmatic, knowing what he likes. More than once, ’s (1798–1861) intensely moving Edo period double-öban composition depicting the seppuku (ritual disembowelling) of Lord En’ya came up in conversation as a personal favourite from the collection (see Fig. 7). At the same time, he ‘love[s] the format of the hashira-e.’ For Te Papa, Ian recommends, not without a little humour: ‘If I had unlimited funds I’d get more hashira; I’d seek better- quality ones that still retain their colour. I consider myself very fortunate to have had five of them in the first place.’13

A collection overview

The core section of this article provides an overview of the diverse sections of the Heriot collection. The authors have selected approximately one-third of Te Papa’s acquisitions to help convey some sense of its scope and quality. Perfect classification is not easy, and the overview is accordingly divided into subject categories that offer a convenient framework for the observations that follow. Respectively, they comprise bijin-ga, kabuki-e, fukeiga, meishö-e, kachöga themes, genre subjects, images from history and literature, surimono, illustrated books and, finally, a small number of what might be termed ‘curios’, two sumi-e and a small silk painting, all of animals.

Bijin-ga

Bijin-ga, ‘beautiful women pictures’, were among the most popular categories of the ukiyo-e repertoire. Twelve works in the Heriot collection illustrate the historical development of bijin-ga, from the contrived elegance of the late eighteenth century to the sentimentality of Meiji period illustration. Harunobu’s Ide no Tamagawa: Yamashiro no meishö (The Crystal River at Ide: A famous spot in Yamashiro) (1769– 70), illustrates the flowering of nishiki-e ‘brocade print’ polychrome printing during the 1760s, and the adoption of the tall hashira-e format that enhanced the floating, lissome elegance of his figures. Kitagawa Utamaro’s (1753– 1806) mannered portrait of the oiran ‘courtesan’ Ichikawa (1790–1800) sustains Harunobu’s chromatic sensibility, but exaggerates the figure’s sensuous elongation beyond

Fig. 2 (right) Two Women Reading, c. 1780–90, hashira-e (pillar print) colour woodblock print, 905 x 250 mm. Artist Isoda Koryüsai (Te Papa 2016-0008-23). 54 Tuhinga, Number 29 (2018)

anatomic reason. Just under a century later, these works container with loquat tendrils and a stylised cloud (kumogata) provided irresistible models for the decadent and grotesque balance the mirroring forms of the two women below. The creations of Aubrey Beardsley (1872–98).14 contrast between the women’s overlapping heads, shoulders Figures of the 1830s by Keisai Eisen (1790–1848) and hems, and the striking negative spaces between their celebrate the decorative in arrangements of opulent heads, shoulders and lower bodies, constructs a deliciously ornament and the adoption of the aizuri-e, ‘blue-favouring energising tension between them. The delicately contrived picture’, trend of the time. By the mid-nineteenth century, juxtapositions emphasise both sensual and poignant Utagawa Sadahide’s (1807–73) Shadow of a Smoker (1850– sensibilities that Koryüsai’s viewers could readily relate to 60) presented viewers with more refined decoration and the iki ‘provocative-chic’ ideals of the licensed pleasure believable figure arrangements. Later Meiji period shin- quarters. Such two-figure constructions particularly hanga ‘new print’ revival bijin-ga by Mizuno Toshikata appealed to Koryüsai. He arranged his figures side by side, (1866–1908) and Miyagawa Shuntei (1873–1914) represent in climbing arrangements of alternating diagonal edges or women and children in domestic or recreational settings in even stacking them vertically in the composition.17 tour de force demonstrations of the printer’s craft and take The contemporary popularity of Koryüsai’s hashira-e full advantage of the synthetic dyes that were then available. was acknowledged by Saitö Gesshin (1804–78) in his Zöho Four prints illustrate key stages in these developments. ukiyo-e ruikö (Enlarged history of ukiyo-e prints) (1844).18 The Koryüsai drew on the pictorial innovations of Harunobu, inventive potentials of the form underpinned this popularity, embracing his soft colour and polychrome decoration, and but more subtle forces were also at play here. Hashira-e were his fondness for the word and picture collisions of mitate-e often mounted as kakemono scrolls, adorned with the artists’ ‘parody pictures’.15 Koryüsai was fascinated by the creative seals used for paintings, and hung in tokonoma ‘display challenges of the difficult hashira-e format, designing more alcoves’. As ‘inexpensive substitutes for kakemono’, they than 450 prints – nearly 40% of all extant works of this offered owners ‘an experience of visual culture normally type.16 He adopted principles for arranging figures and reserved for the higher classes or the wealthy’.19 Koryüsai’s minimal contextualising features (landscape references, subtle glimpses of ankles, wrists or a décolletage introduced architecture, branches or clouds) that have their sources in a voyeuristic sense of intimacy, enhanced by the way the Chinese and (794–1185) kakemono ‘vertical narrow format suggested a furtive glimpse through a partly scroll’ painting precedents. In Fig. 2, a hanging plant opened sliding door or screen.20 These qualities drew

Fig. 3 Three Women Viewing Cherry Blossom on a Riverbank, c. 1850, three-öban-sheet colour woodblock print triptych, 567 x 1074 mm (total). Artist Utagawa Kunisada (Te Papa 2016-0008-33). Rising sun at Te Papa: the Heriot collection of Japanese art 55

his viewers into playful engagements, encouraging their themes of the ephemerality of beauty and youth that inform constructions of imaginary narratives around his pictorial a sensibility of mono no aware, a sensitivity to the pathos suggestions.21 of things. The enclosed pleasure craft in the background If Koryüsai capitalised on suggestion, intimacy or reminded viewers of the riverside sex trade and the fleeting imaginative engagement, the bijin-ga of Kunisada (Fig. 3) nature of pleasure. The triptych format encouraged viewers adopt an altogether more pragmatic attitude. Kunisada’s to make comparisons between these well-known women of three women are yüjo, ‘women for play’, from Edo’s the highly competitive world of the pleasure quarters, while licensed pleasure quarters, engaged in a hanami (daytime promoting trends in fashion and textile design. Although it viewing) of sakura (cherry blossom). The association risks being anachronistic and Eurocentric to call this triptych between beautiful young women and sakura resonated ‘Baudelairian’, the eponymous French critic, writing not

Fig. 4 Yuagari: Kansei koro fujin (After the Bath: Woman of the Kansei era (1789–1801)), no. 4 in the series Sanjü-rokukasen (Thirty-six Types of Beauty), published from 1891 (1893), nishiki-e colour woodblock print, 371 x 247 mm. Artist Toshikata Mizuno (Te Papa 2016-0008-50). 56 Tuhinga, Number 29 (2018)

long after the time it was printed but located a world away, protection, reinforces the viewer’s sense of familiarity here. would surely have admired Kunisada’s ability to recapture Toshikata’s intimate scenes provide evocative, atmospheric ‘the circumstantial element’ of beauty, embodying ‘the age, contexts for representations of pensive figures.26 His Kansei its fashions, its morals, its emotions’.22 era (1789–1801) woman seems quietly melancholic, lost in Toshikata Mizuno’s (1866–1908) bijin-ga reflect changes her own thoughts in a moment of private reverie. in the status and craft of the woodblock print during the The delicacy of Toshikata’s romantic figures demanded Meiji era. The survival of the medium through this period the most refined craftsmanship. This technical tour de is surprising; its technical capacities were surpassed by force also enriches the prints of Shuntei (Fig. 5) in bokashi those of photography, lithography and photogravure. The colour gradations, applications of mica, metallic pigments woodblock retained important roles in the publication and shomen-zuri burnishing. The collection contains three industry, however. Its accommodation of colour and gorgeous triptych designs by Shuntei, each a composition representational licence allowed Toshikata to compose from his 1898 series Bijin junikagetsu (Beautiful Women for dramatic celebrations of Japanese successes in the First the Twelve Months), representing January to February, Sino-Japanese War (1894–95),23 for example, and to design March and October. Like Toshikata, Shuntei spent periods delicate kuchi-e frontispiece (literally ‘mouth picture’) training in the studios of ukiyo-e designers, and his principal and sashi-e ‘inserted illustrations’ for romantic novels and occupation was as an illustrator for newspapers, magazines newspapers like Yamato Shimbun.24 Where Kunisada’s works and books. His delicate subjects are set in evocative seem detached, Toshikata’s illustrations are sentimental atmospheric landscapes, and their women and children and a little ethereal. The former quality reflects the tenor reflect a meeting of the bijin-ga fashion statements of the of contemporary popular literature, the latter the fluid past, and the family values and accoutrements (such as the transparency of Satsuma porcelain painting, in which umbrella illustrated in Fig. 5) of a Western-style modernity Toshikata had trained. His works also maintain sensibilities in Japan.27 In some ways, these compositions draw on from earlier eras.25 The casual intimacy of his view of a Yamato-e ‘Japanese picture’ genre traditions, but their focus woman after the bath (Fig. 4) echoes the provocative chic on family groups and leisure activities offers delightful views sensibility of the Edo ethos of iki. The presence of the of late Meiji optimism. cat, celebrated in Japan as an emblem of good luck and These works also juxtapose the beauty of nature with

Fig. 5 Bijin junikagetsu (Beautiful Women for the Twelve Months): Number 10 – Koyo (October), 1898, three-öban-sheet colour woodblock print triptych, 376 x 717 mm (total). Artist Miyagawa Shuntei (Te Papa 2016-0008-49). Rising sun at Te Papa: the Heriot collection of Japanese art 57

the fragile beauty of women and children. Their subjects of special importance in the kabuki repertoire: Kanadehon draw on pastimes of garden viewing that date back to the chüshingura (The treasury of loyal retainers) (1748). Chüshingura Heian era. In doing so they sustain a sensitivity to seasonal is a vendetta tale. Its popularity drew on its origins in real transience that evoked nostalgic aesthetic sensibilities. The historical events that culminated in the vendetta of the 47 reverence for nature evident in these works appealed equally rönin in 1703, and its values of loyalty and honour. Its 11 to early New Zealand visitors to Japan: Isabella and Frederic acts accommodate the theatrical gamut, from heart-rending Truby King collected such works during their visit in love stories, to grisly murders, earthy sub-plots and bloody 1904.28 This popularity contrasts with Shuntei’s reputation battles.29 in Japan, where his works were enjoyed as ephemera, but Narrative representations of Chüshingura usually followed then ignored and even lost. Only recently, certainly well the theatrical structure closely, each print representing beyond the artist’s brief lifetime, has their status within a key scene from a particular act of the play. Characters Japan appreciably risen, thanks to the attention of ‘offshore’ were generally arranged in the highly expressive mie, or scholars such as Helen Merritt and Julia Meech-Pekarik. ‘frozen poses’, that punctuated each scene. Although they illustrated only a single dramatic moment, these images Kabuki-e offered sufficient information for knowledgeable fans – and The popular art form of printmaking found a comfortable there were many – to reconstruct the surrounding events in commercial partner in kabuki theatre. Kabuki-e theatre their imaginations. The opening act of the play relates the pictures promoted productions in the highly competitive events leading to the humiliation of a young En’ya Hangan theatre world, celebrated the actors and characters of kabuki (historically Asano Naganori, Lord of Akö, 1667–1701) by narratives, and reconstructed their Byzantine plots in serial the köke ‘high-house’ court official Kö no Moronao (Kira print sequences. The Heriot collection includes 12 prints Yoshinaka, 1641–1703). The subsequent episodes reveal the with kabuki-related themes: seven portraits of actors in role narrative that then unfolded. and five narrative scenes. They include works by the Osaka The first of four prints by Eisen represents a key moment artist Ryüsai Shigeharu (1802–53), and Edo artists Kunisada, in Act II, ‘Momonoi no yashiki’ (‘At Momonoi’s mansion’) Kuniyoshi, Eisen, Sadanobu Hasegawa III (1881–1963) and (Fig. 6). Moronao’s slight had been severe; Momonoi Shunsen Natori (1886–1960). Six prints illustrate one play Wakasanosuke, on the right-hand side of the composition,

Fig. 6 Kanadehon Chüshingura. Act II: Momonoi no yashiki: matsu-kiri (At Momonoi’s Mansion: Cutting the pine), c. 1820, öban colour woodblock print, 282 x 360 mm. Artist Keisai Eisen (Te Papa 2016-0008-2) 58 Tuhinga, Number 29 (2018)

Fig. 7 Kanadehon Chüshingura. Act IV: Seppuku of Lord En’ya, 1850, double-öban colour woodblock print, 500 x 595 mm. Artist Utagawa Kuniyoshi (Te Papa 2016-0008-37).

has vowed to avenge his friend. At far left we see Öboshi the senior official. Hangan is sentenced to take his own Rikiya, son of the play’s hero, Öboshi Yuranosuke, arriving life by performing seppuku. Kuniyoshi’s illustration for Act with a message from Hangan to Wakasanosuke. He and IV focuses on the final moments of Hangan’s life (Fig. 7). his fiancée, Konami, watch as Wakasanosuke’s retainer, The setting is his own home – identifiable by the striking Honzö, pledges support for his master, seizing his sword and mon (emblems) of crossed feathers on a blue field. At right dramatically slashing a branch of pine. He replaces the blade are the kneeling figures of the shogun’s envoys, Ishidö and in its scabbard uncleaned, so that the sap on the blade would Yakushiji. In the rear, Hangan’s retainers Hara Göemon prevent Momonoi drawing it in anger and shaming himself. and Ono Kudayü witness the event. You can hear a pin The other Eisen compositions illustrate Act III, ‘Kamakura drop. Hangan kneels left of centre, on a white quilt, with denchu Ashikaga-yakata monzen-no-ba’ (‘Before the gate of sakura at each corner. His chief retainer, Yuranosuke, the Ashikaga Palace at Kamakura’); Act VI, ‘Kanpei sumika’ receives his master’s dying wish: to avenge the wrong that (‘At the dwelling of Kanpei’); and Act VIII, ‘Tökaidöō has been done. Kuniyoshi’s claustrophobic composition michiyuki’ (‘The journey along the Tökaidö’) is tightly grouped. Its sober arrangements of white, umber, In Act III, Wakasanosuke and Hangan are invited to blue and black, and its refined printing, including shomen- festivities at the Ashikaga Palace in Kamakura. Hangan, zuri ‘polished ink’ details, compellingly emphasise the again goaded by Moronao, draws his sword and attacks gravity of the moment. The stage is set for the narrative Rising sun at Te Papa: the Heriot collection of Japanese art 59

Fig. 8 Actor in the Role of Hayano Kanpei in the Kanadehon Chüshingura, 1854, öban colour woodblock print, 535 x 380 mm. Artist Utagawa Kunisada (Te Papa 2016-0008-30). in which Yuranosuke leads 46 of Hangan’s loyal retainers his father-in-law, Yoichibei, to sell his daughter Okaru to towards the final, violent enactment of their vendetta. a Gion brothel to raise money to support the vendetta. In Kunisada’s portrait of the retainer Hayano Kanpei this scene from Act V, he is seen just at the moment when (Fig. 8) presents the character frozen in a dramatic mie he thinks he has shot a man while hunting in the forest. pose, feet heel to heel, huddled under the rain, and looking Later, believing he has killed Yoichibei, he too commits wistfully to his left. His attitude is significant; Kanpei seppuku just as his honour is reconfirmed.Chüshingura occupies an ambivalent place in the Chüshingura narrative. was so popular that Kunisada’s viewers would have had His neglect of his duties in Act II was a contributing factor little difficulty in recognising the reasons for Kanpei’s to Hangan’s fate, and he flew to the countryside in shame. quizzical gaze. Later, he vowed to support the loyal league, allowing Natori’s portrait of the actor Morita Kan’ya XIV (1907– 60 Tuhinga, Number 29 (2018)

75) illustrates two further characteristics of ukiyo-e theatre prints (Fig. 9). First, it is an okubi-e, or ‘big-head picture’. The close-up format allowed artists to design recognisable portraits of actors in key roles or expressive poses. Second, artists of the early twentieth-century shin-hanga revival retained the services of skilled block cutters and printers, and redefined the stylistic tropes of the genre. Natori maintains the Japanese-style idioms that he had learned while studying Nihonga, ‘Japanese painting’, under Kubota Beisen (1852–1906), Kubota Kinsen (1875–1954) and Hirafuku Hyakusui (1877–1933).30 His shin nigao-e, ‘new actor portraits’, in the okubi-e format are recognisable but finely stylised, their simply delineated subjects positioned against plain, sometimes textured, backgrounds.31 Natori’s prints were popular with his Japanese public, but they reflected another departure in Japanese print publication, commanding a parallel appeal in the markets of North America.

Fukeiga

North American viewers were equally receptive to shin- hanga fukeiga, ‘landscape compositions’, especially after the seminal Toledo Museum of Art exhibitions of the 1930s.32 The Heriot collection includes nine shin-hanga landscapes, Fig. 9 Morita Kanya as Kitsune Tadanobu in the Play by Narazaki Eishö (1864–1936), Kasamatsu Shirö (1898– Yoshitsune senbonzakura (Yoshitsune and the Thousand Cherry 1991), Hiroshi Yoshida (1876–1950), Takahashi Hiroaki Trees), 1952, öban colour woodblock print, 392 x 273 mm. (Shötei, 1871–1945), Bannai Kökan (1900–63) and Hasui. Artist Shunsen Natori (Te Papa 2016-0008-46). These complement Te Papa’s holdings of landscape prints by Hokusai, Andö Hiroshige and Hiroshige II (Shigenobu, 1826–69). The quietly evocative seasonal themes of these later Meiji and Taishö era landscape views reflect Ian Heriot’s personal preferences, especially as they developed in the later years of his collecting. They presented opportunities to acquire prints in fine condition, and their less theatrical ambience seems consistent with the attitudes of humility that suffuse Heriot’s own account of his collecting.33 These works reflect the tastes for fine craftsmanship, naturalistic composition and nostalgia for timeless rural themes that emerged during these eras. They also embrace Western devices of chiaroscuro and atmospheric and linear perspective in naturalistic compositions that appealed to the Japanese and Western public alike. Rising sun at Te Papa: the Heriot collection of Japanese art 61

Fig. 10 Asakusa Sensöōji daichöchin (The Great Lantern of the Sensö Temple, Asakusa), 1934, öban colour woodblock print, 385 x 265 mm. Artist Kasamatsu Shirö (Te Papa 2016-0008-19).

Temples and shrines were popular subjects for shin- meishö-e, together with votive prints and other souvenirs. hanga artists. These prints maintained traditions of Although Shirö had studied Nihonga under Kaburagi Kiyokata meishö-e and were often acquired by visitors to well- (1878–1972), and also woodblock printing,34 this print reveals known centres. Sensö-ji (Asakusa Kannon Temple) his wider interest in Western-style art in his adoptions was exceptionally popular, attracting huge crowds, of linear and atmospheric perspective, diminutions of scale, especially during the new year. Shirö’s view (Fig. 10) and effects of light to describe the dimly lit ambience of the focuses on its shop, where visitors could purchase such interior. 62 Tuhinga, Number 29 (2018)

Fig. 11 Söchüū Maekawa no ame (Rain in Maekawa, Söchüū), from the series Tökaidöō fükei senshü (Selection of Views of the Tökaidö), 1932, öban colour woodblock print, 388 x 262 mm. Artist Hasui Kawase (Te Papa 2016-008-11).

Hasui’s atmospheric fukeiga achieved international of Hasui’s works is enhanced by his depiction of nostalgic popularity. His poignant view of a figure walking at night subjects in rural settings, often represented at night, dawn in pouring rain (Fig. 11) was one of his own favourite or sunset. Underpinning these qualities are five conceptual works.35 In Hasui’s own words, ‘I love quiet scenes with the foundations: a belief in the inspirational potential of nature; shadow of loneliness hanging over them’.36 He emphasised an appreciation of the power of atmospheric effects to that sense of solitude by manipulating relations of scale and suggest ephemerality; the importance of poetic moods of pictorial depth, and through the subtle bokashi modulations ‘tranquillity or contemplation’; the source of a Japanese of light and colour he had learned through his study of sense of identity in values and sensibilities of the past; and the the kösen-ga, ‘light pictures’, of his teacher Kobayashi capacity of ‘beauty’, especially that of nature, for ‘expressing Kiyochika (1847–1915).37 The lyrical sense of loneliness aesthetic, social, philosophical and even spiritual beliefs’.38 Rising sun at Te Papa: the Heriot collection of Japanese art 63

Hasui realised these concepts in quiet constructions of ordinary, unremarkable scenes39 that juxtapose a sense of transience and melancholy with timeless qualities of tranquillity and peace.40 Kachöga A reverence for nature also found a sympathetic medium in kachöga or kachö-e, ‘bird and flower pictures’. Kachöga embraces the wider natural history field of insects, fish and larger animals.41 The genre has a long pedigree in Japan, maintaining metaphoric associations – the crane (tsuru) signifying longevity, for example, and the swallow (tsubame) suggesting good fortune.42 Though these associations survive today, twentieth-century artists favoured these subjects because they afforded intimate insights into the profound beauty of nature. The Heriot collection includes seven kachöga. Five bird compositions by Ohara Koson (also signing as Shöson and Hoson; 1877–1945), published by Watanabe Shözabüro between 1927 and 1937, include images of egrets, an eagle, a cockatoo and a pair of godwits. Two flower pictures by Kawarazaki Shödö (1889–1973) were published by Unsödöō in the mid-1950s. The Koson compositions are

Fig. 13 Peonies, 1955, öban colour woodblock print, 400 x 270 mm. Artist Kawarazaki Shödö (Te Papa 2016-0008-21).

surprisingly varied – at turns delicately elegant, majestic or whimsical. An intimate view of the godwits wading in front of gently windblown reeds exemplifies the artist’s painterly style, and the gentle naturalism of his kachöga (Fig. 12). The painterly quality is evident in this work in the mogu, ‘boneless’ technique (i.e. without outlines), the single, fluid brushstroke seen here in each leaf in the rushes, and the rippling water. Together with the transparency of his pigment and muted palette, these qualities draw on Koson’s study of Maruyama-Shijö school painting under Suzuki Kason (1860–1919).43 Koson began designing kachöga early in his career and completed more than 450 bird compositions.44 The majority represented familiar species like the godwits shown here – humble subjects that encouraged a quiet, introspective engagement from his viewers. Their carefully contrived naturalism45 and painterly effects demanded the finest craftsmanship of his printers. The combination was appealing, and his works found international popularity. Fig. 12 Two Bar-tailed Godwits in Shallow Waters Next to Where Koson’s works were soft, painterly and Reeds, 1926, öban colour woodblock print, 395 x 265 mm. atmospheric, the flower designs of Shödö were crisp, sharply Artist Ohara Koson (Te Papa 2016-0008-28). delineated and botanically correct (Fig. 13).46 In some ways, 64 Tuhinga, Number 29 (2018)

Fig. 14 Shin yoshiwara sakura no keshiki gomai tsuzuki (Cherry Blossoms in the New Yoshiwara), 1811, öban nishiki-e woodblock print, 503 x 371 mm. Artist Utagawa Toyokuni (Te Papa 2016-0008-53). their fine observation drew on kachöga precedents; earlier enjoyed for their nigao-e ‘true-view’ realism. The springtime flower prints often had their botanical detail clearly labelled setting of the rightmost print of a pentaptych by Toyokuni like textbook diagrams. In Shödö’s designs, however, their (Fig. 14) suggests that the other panels recorded the carefully precision is complemented by a decorative linear elegance. choreographed procession of the oiran dochu. The dochu was Barely a decade later, but in a very different cosmos, finely performed four times during the year, to coincide with the wrought idioms like these would be adopted by the flower- change of the seasons. It offered the women of the Yoshiwara power poster designers of Haight-Ashbury. licensed quarters a rare opportunity to move beyond their cloistered worlds, and the public a chance to view beauties Genre subjects much celebrated in bijin-ga prints. In this instance, oiran courtesans, their ‘younger sister’ kamuro and yüjo prostitutes Ukiyo-e artists drew on the conventions of Yamato-e scenes of are jostled by touts, a delivery boy and norimono ‘palanquin’ daily life; in early commentaries, ukiyo-e was defined largely carriers as they pass through Yoshiwara Ömon, the Great in terms of these preoccupations. A small group of works in Gate of Yoshiwara. Toyokuni’s viewers would have delighted the Heriot collection represent the human fabric of Edo. in the carefully observed variety of dress, pose and gesture of Utagawa Toyokuni’s (1769–1825) representations were much each character in the composition. Rising sun at Te Papa: the Heriot collection of Japanese art 65

Fig. 15 Tango no sekku (Boys’ Festival Procession Across the Nihonbashi Bridge), 1843–45, three-öban-sheet colour woodblock print triptych, 361 x 771 mm (total). Artist Utagawa Kuninao (Te Papa 2016-0008-29).

Festivals, whether celebrating the seasons, anniversaries of temples and shrines, or other significant events, provided entertaining subjects. Tango no sekku, the Boys’ Festival, was long celebrated on 5 May to pray for the health and courage of boys. Today, although its commemoration is embraced in Kodomo no hi, Children’s Day, boys still often enjoy special favour. Continuing in the nigao-e tradition, the little-known Utagawa Kuninao (1793–1854) adopts an öban triptych format to capture the dynamic force of a crowd of ebullient boys rushing across the Great Bridge of Japan (Fig. 15). The different boys’ clubs are indicated by their distinctive happi jackets and decorative lanterns. More inventively, however, Kuninao has managed to describe a different facial expression for every single figure, while conveying the headlong, careering excitement of all. Viewers would have laughed outright at the sight of Utagawa Hirokage’s (active 1855–65) hapless group (Fig. 16), delighting in the slapstick humour of their ineffectual responses to the gusts that have thrown them – and their possessions – into disarray. It could be said that this is a particularly pertinent choice for a Wellington collection! On looking closely, however, viewers might have found a more nuanced pleasure as they noticed that Hirokage’s composition was a pastiche of one of his master Hiroshige’s famous views of Edo, and the flying objects have in turn Fig. 16 Shin-Öhashi no ökaze (Strong Wind on Shinhashi been purloined from compositions by Hokusai. Although Bridge), from the series Edo meishöō döke zukushi Hirokage may not be the most original of artists, and the (Comical Views of Famous Places in Edo), 1859, öban rendition of detail here seems clumsy, works like these offer colour woodblock print, 367 x 246 mm. Artist Utagawa today’s viewers more immediate gratification than earlier, Hirokage (Te Papa 2016-0008-13). 66 Tuhinga, Number 29 (2018)

canonical Edo constructions ever allowed. In turn, viewers River she found the figure of her son. Overjoyed, she tried of the period could surely see themselves in any of these to say something to him; then, the figure turned into a works, appreciate the specific situations described in each, willow tree.’47 Utagawa Yoshiiku’s (1833–1904) view of the and recognise the social crush of this, the busiest and most figure of Hanako and Umewakamaru at the moment of populous city in Japan. his transformation reflects a widely held ‘belief in lingering spirits’48 caught in a suspended state between the world in Images from history and literature which we live (kono-yo) and the world beyond (ano-yo). This liminality, evident in Japan in the Jizö motif, recognises a state Complex threads of superstition and supernatural in which ‘the soul is perceived not as transitory; death even phenomena provided popular motifs for artists and the Edo constitutes a kind of aesthetic, in spite of one’s abhorrence populace. The narrative in the cartouche at the top of the of the process’.49 That sensitivity suffuses the Heian narrative composition seen in Fig. 17 explains the pictorial scene: of Genji monogatari (The tale of Genji). The yellow cartouche ‘Hanako is the wife of Major General Yoshida Korefusa. She at upper right in Yoshiiku’s print carries the series title and chased her lost son Umewakamaru, wandering around in characters for chapter 39, ‘Yügiri’ (‘Evening mist’), of Genji eastern countries [i.e. the Kanto region] and by the Sumida monogatari. Curiously, however, the Genji-mon (chapter logo) on the green ‘page’ behind this cartouche are for chapter 21, ‘Otome’ (‘The maiden’). Unsurprisingly, neither chapter refers to the narrative included in Yoshiiku’s work – the respected calligrapher and senior court minister Fujiwara no Korefusa (1030–96) in the tale was born two decades after the original publication of the novel. Two Kunisada bijin-ga illustrate the links that Edo chönin, ‘townsmen’, maintained with the classical past of the Heian court. In each, a small portrait of a poet and lines of verse on two playing cards at upper left refer to the game popular in middle-class domestic settings and brothel worlds alike, where players match halves of poems. In the example shown in Fig. 18, the short waka verses are from the famous Heian collection Hyakunin isshu (One hundred poems by one hundred poets). The scroll cartouche at upper right contains a biographical note about the poet Chünagon Ötomo no Yakamochi (718–785), editor of the Man’yöōShüū verse collection (c. 759) and one of the Thirty- six Poetic Immortals. The poem reads:

kasasagi no wataseru hashi ni oku shimo no shiroku wo mireba yo zo fukenikeru When I see the whiteness

Fig. 17 Yoshida no hanako (Hanako, Yoshida’s Wife), Chapter of the frost that lies 39 Yügiri (Evening Mist), from the series Imayö nazorae Genji on the bridge the magpies spread, (Modern Parodies of Genji), 1864, öban colour woodblock then do I know, indeed, print, 358 x 232 mm. Artist Utagawa Yoshiiku (Te Papa 2016-0008-58). that the night has deepened50 Rising sun at Te Papa: the Heriot collection of Japanese art 67

Fig. 18 (left) Poet no. 6: Chünagon (Ötomo no) Yakamochi, from the series Hyakunin isshu (One Hundred poems by one hundred poets), c. 1844, öban colour woodblock print, 378 x 259 mm. Artist Utagawa Kunisada (Te Papa 2016- 0008-34).

The juxtaposition of a classical verse with the image of a yüjo brothel-quarter woman throughout this series may seem incongruous, but women of the Yoshiwara were expected to have a knowledge of classical verse and to embrace it into the verbal intercourse of their engagements. During the nineteenth century they also adopted other Heian habits, particularly in mannerisms of dress and deportment. Kunisada’s fellow Utagawa school student Kuniyoshi was as versatile as his contemporary. Subsequently, however, Kuniyoshi’s warrior scenes from historical narratives like Heike monogatari ()51 have received the closest attention. His dramatic Ujigawa gassen no zue woodcut (Fig. 19) depicts events early on New Year’s Day 1184 in the second battle of Uji River (near ) during the (1180–85) between the Minamoto (Genji) and Heike (Taira) clans. The Heike victor in the first battle of Uji River had been the rebel Minamoto leader Yoshinaka (1154–84), fighting against his own clan members. Here, four years later, the Minamoto forces strike back. Minamoto Lord Yoritomo (1147–99) has gifted two steeds, the pure-

Fig. 19 Ujigawa gassen no zue (The Battle at Uji River), 1849, three-öban-sheet colour woodblock print triptych, 1104 x 368 mm (total). Artist Utagawa Kuniyoshi (Te Papa 2016-0008-38). 68 Tuhinga, Number 29 (2018)

black horse Surusumi (Ink-stick) to the general Kajiwara However, his historical series, musha burui Genda Kagesue (1162–1200), and the chestnut Ikezuki (Ill- (Yoshitoshi’s Courageous Warriors; alternatively Yoshitoshi’s tempered Biter) to Sasaki Shirö Takatsuna (1160–1214). Warriors Trembling with Courage)53 offered his viewers a more Takatsuna rashly claimed he would be first across the mature reflection on the military heroes of the past. Thirty- river.52 Here, Kagesue is depicted on the left-hand sheet; three designs were published by Kobayashi Tetsujirö Takatsuna, pictured oddly with a dappled grey steed, is between December 1883 and December 1886.54 The in the centre; and Hatakeyama Shigetada (1164–1205) is Heriot collection includes portraits of the military heroes at the right. Takatsuna reaches the further shore first, and Chinzei Hachirö Tametomo (1139–70), Hino Kumakawa- he and Kagesue are followed by Shigetada and 500 men. maru (fourteenth century) and Taira (Sökoku) no Kiyomori The arrow flying low over the water on the right-hand (1118–81) (Fig. 20). sheet had been released from the far side by Yamada no Kiyomori was the leading figure in the Heike monogatari. Jirö, and buried itself in the brow of Shigetada’s horse. The Following the death of his father in 1153, he assumed Minamoto victory that followed proved a turning point in leadership of the . Kiyomori enjoyed a rapid the wars, and a decisive moment in the rise to power of the court promotion after his victory over Minamoto forces in mercurial young general (1159– the 1159 Rebellion, becoming DaijöōDaijin – chief 89). government minister – in 1167. He died during the first The mid-career works of Kuniyoshi’s most significant years of the Genpei Wars. In 1168, Kiyomori built the student, Tsukioka Yoshitoshi (1839–92), were known for Itsukushima Shrine on the waterfront at Miyajima Island. compositions representing bloody violence and cruelty. Yoshitoshi has depicted him rather imperiously here, arm raised high against the , brow furrowed in concentration, as he fights to hold back the setting of the sun in order to keep his promise to complete work on the shrine before sunset. His authority is emphasised by his stature, dominating the height of the composition beside the small figure of a court maiden. The striking folds of his black-and-white robes contrast with the velvet-like surfaces of intense red as he stands at the water’s edge. The intensity of the foreground colour contrasts again with the softer bokashi ‘gradations’ of red, through orange and yellow, into the chill-blue evening sky. Yoshitoshi’s master printers have brought out an astonishing woodgrain effect, which appears to emerge like fiery rays from of the sun itself. Kiyomori’s Taira family shrine at Itsukushima and its famous ‘floating’ torii gate are popular tourist attractions in Japan today.

Surimono Ian Heriot maintains a particular interest in the privately commissioned, finely printed surimono ‘rubbed thing’ print format.55 It is easy to see why. For their cultivated patrons, surimono informed insights into the Japanese aesthetic and poetic consciousness, resonating Heian court tastes56 and constructing an elegance of craft in subtleties of colour and linear finesse. Thus, the nuanced tonalities of Ryüryükyo Shinsai’s (active 1799–1823) Two Birds and the Moon evokes chill themes of mono no aware melancholy. The delicate Fig. 20 , no. 18 in the series Yoshitoshi mistiness of Totoya Hokkei’s (1780–1850) Blue Landscape musha burui (Yoshitoshi’s Courageous Warriors), 1883–86 draws on Chinese tastes for monochrome suggestion. In (mid-1885), öban colour woodblock print, 358 x 240 mm. contrast, Hokkei’s earthy depiction of Two Men Dancing Artist Tsukioka Yoshitoshi (Te Papa 2016-0008-61). Next to a Geisha Playing a Samisen is hilarious, and Utagawa Rising sun at Te Papa: the Heriot collection of Japanese art 69

Toyohiro (1773–1829) brings a subtly wrought wit to his The first poem mentions the legend of Rosei, a poor juxtaposition of the deities Ebisu and Benten. Chinese traveller who fell asleep while his meal was being Toyohiro’s surimono (Fig. 21) was composed as a prepared at an inn in the village of Kantan, in China. celebration of the New Year. The ox-cart provided a Before he was served his meal, he dreamed of a lifetime popular mode of transport for the Heian court aristocracy – of happiness, honour and prosperity. In the second poem, their kichö (curtains of state) or sudare (rolled reed blinds) and ‘the mother of the flowers’ is the rain. There is a pun on decorative swathes of ribbon allowed women to be conveyed oru, ‘to break’, oriha, ‘to alight’, and oriori, ‘sometimes’. 57 in anonymity. The ox-cart and the kumogata mist across the Kantan no makura ni aranu yume samite eiga no hana ni top of the composition reminded Edo viewers of classical niou ganjitsu literature like Genji monogatari or the great collections of verse that inspired so many surimono kyöka, or elegant ‘wild It was not until the dream of the pillow at Kantan, when verse’ surimono.58 From Heian times, and well into the Edo I awoke from my dream on New Year’s Day, I smelled the period, the Genji and elegantly decorated ox-cart associations flowers of prosperity Hana no Harundo were synonymous with refined sensibilities of femininity. Waga yado no ume wa oranedo oriha ni wa oriori konomu Toyohiro’s Edo public would have been familiar with these themes, whose significance had been much revived through hana no kakasama the Tokugawa nineteenth-century kokugaku, ‘national Although she does not pluck the blossoms of my plum study’, initiative.59 Kokugaku scholars sought to reinvigorate tree, the mother of the flowers alights to enjoy herself Japanese study of the national classics in verse, prose and sometimes Sakuragawa Jihinari60 historical texts. This informed the sophisticated historical and literary knowledge that educated chönin could bring to An early source of surimono lay in the production of their appreciations of works like this. Roger Keyes explains egoyomi ‘calendar prints’. Hokusai’s delectably enigmatic the context of the elegantly brushed kyöka at left in this composition celebrating 1822, the year of the horse, is composition, and provides translations of each: seen in Fig. 22. The tobacco-pouch clasp, printed in silver

Fig. 21 Court Carriage, c. mid-1810s, surimono woodblock print, 199 x 185 mm. Artist Utagawa Toyohiro (Te Papa Fig. 22 Uma no senbetsu (Farewell Gift for the Horse), 2016-0008-51). from the series Uma zukushi (A Selection of Horses), 1822 (second-edition ‘A-grade’ facsimile, c. 1880s), shikishiban (blind-printed) surimono woodblock print, 201 x 180 mm. Artist Katsushika Hokusai (Te Papa 2016-0008-16). 70 Tuhinga, Number 29 (2018)

pigment in the stylised shape of a reclining horse, confirms the horse theme. The attached pipe case leans on a round box of seasoned fish (denbu). It is labelled ‘Umagae Dembu Hasegawa- chö, Togokuya’, the name and address of the shop where it was sold. In the foreground is a small cloth case printed with shikishiban embossing. A toothpick and its carrying case rest on the box.61 Matthi Forrer describes the simple arrangement of Shözo Ittei’s (active 1820–30) charming surimono seen in Fig. 23.62 He identifies the bowl’s contents as dandelion leaves, and the textured battledore-shaped implement as a grater. The characters on the bowl are those for fuku (luck) and fu (long life), while the flowing silver emblem is that of the poetry club that assembled around Tsurunoya Osamaru (c. 1751–c. 1839). The cloth bag in the form of a monkey attached to the scissors indicates 1824, the year of the monkey. The poems are by Ki no Takane (n.d.), Numakane Jun (n.d.) and Tsurunoya. The seal at the lower left identifies the publisher Tani SeiködöōTakuboku (active 1822–31). Tani Seiködö is unique in applying his own seal to almost every print he produced Fig. 23 Porcelain Bowl and Scissors, Issued by Followers of during his prolific decade of work from 1822.63 He was remarkable the Poet Tsurunoya Osamaru, 1824, nishiki-e woodblock print with metallic pigments and for his assumption of the roles of block-cutter and printer, the shikishiban (blind printing), 236 x 211 mm. Artist refinement he brought to both crafts, and for his close relationships Shözo Ittei (Te Papa 2016-0008-18). with the literary circles of his time.64

One of the most elegant of the Heriot surimono, refined and simple in its means, is Shibata Zeshin’s (1807–91) undated composition of a humble gourd flask, cup and butterfly – a paean to Taoist values of living a humble life, in tune with nature and at one with the world (Fig. 24). The assurance of Zeshin’s work stems from the peculiar synthesis of his early training in lacquer making, studying painting with Suzuki Nanrei (1775–1844) and Okamoto Toyohiko (1773–1845), and studying literature with Kawara Keiju (n.d.) and Rai San’yöō (1780–1832). He also had good knowledge of haiku poetry and the tea ceremony.65 The poetic simplicity of Zeshin’s surimono and drawings was much favoured by the great French collector Henri Vever (1854–1942),66 and their humility found parallel sympathies in the Heriots’ own attitudes to ukiyo-e and art collecting.67 Zeshin melded a reverence for nature, Chinese- style asymmetrical arrangements, and an elegant Fig. 24 Gourd, Butterfly and Basket, mid- to later nineteenth Maruyama-Shijöō school fluidity of gyö, the ‘semi- century, surimono woodblock print, 240 x 250 mm. Artist Shibata Zeshin (Te Papa 2016-0008-63). formal’ cursive brush mark. Rising sun at Te Papa: the Heriot collection of Japanese art 71

Fig. 25 (left) Blind Men Measuring the Elephant, from Hokusai manga. Vol. VI, 1817 (Nagoya, 1878 edition), woodblock-printed album, 226 x 310 mm (double-page spread). Artist Katsushika Hokusai (Te Papa 2016-0008-17).

Fig. 26 (left) Hokusai’s Theories for Composition Using Linear Perspective and a ‘Rule of Three’; African Figures, from Hokusai manga. Vol. III, 1815 (Nagoya, 1878 edition), woodblock- printed album, 226 x 310 mm (double-page spread). Artist Katsushika Hokusai (Te Papa 2016-0008-17).

Illustrated books works first issued in spring 1815 (II, III), summer 1816 (V), 1817 (VI), spring 1819 (VIII) and 1819 (IX).70 The Heriots also developed a fascination for the block- The illustrations of the manga offered Hokusai’s public printed book. Those acquired by Te Papa include six astonishing displays of his virtuoso drawing skills. There volumes of the famous Hokusai manga, which comprise a seems little logic in their arrangements; there is no written vast collection of approximately 4000 sketches of diverse rationale and no sense of ‘homogeneity or over-all pattern’.71 subjects: landscapes, flora, fauna, everyday life, technology If the title of the first volume suggests some instructional and the supernatural world. The term manga certainly did intention, the subsequent issues seem to have been conceived not suggest today’s ‘comic-book’ form, but instead referred simply as vehicles for the pictorial explorations of Hokusai’s to ‘random sketches’ or ‘a multitude of drawings of a lively, own polymathic curiosities, and published as popular sketch type, often presented in numbers on a single page’.68 visual entertainment. His illustration of blind men and an The Hokusai manga were published by Eirakuya Töshirö elephant (Fig. 25) reflects his breadth of interests. It draws in Nagoya, or by Kadomuraya Jinsuke in Edo. The first on precedent images from Japanese artists – by Hanabusa volume was published in early spring 1814 under the title Itchö (1652–1724), for example – but its theme of blind Denshin kaishu (A primer for transmitting the true spirit).69 men touching different parts of an elephant, then disagreeing Those in the Heriot collection are of an 1878 edition of completely on the nature of the beast, originates from the 72 Tuhinga, Number 29 (2018)

Fig. 27 An American Woman Making Herself Up with the Aid of a Double Mirror, from Yokohama kaikö kembun-shi (Things seen and heard at the Yokohama Open Port). Vol. V, 1865, woodblock-illustrated volume, 253 x 350 mm (double- page spread). Artist and author Gyokuransai Hashimoto (aka Utagawa Sadahide) (Te Papa 2016-0008-41).

Indian subcontinent. The theme was known in Europe and Volume 5 in the Heriot collection includes diverse views of America by the nineteenth century72 and survives today. Yokohama’s ‘shanty-town’ streets, warehouses and shops, The immediate success of Hokusai’s manga reflects their and its Western visitors; not always flatteringly, as images of inventive melding of humour and realism with a sense of the boorish foreigners attest.76 ‘common touch’ of his viewers’ worlds.73 His preoccupations Representations of Westerners at work or at home with all things Japanese engaged complementary curiosities provided insights into the ways they re-created their own of his public. An equally random collation of views from worlds in the Japanese settlement.77 Sadahide’s triple portrait abroad – from African figures to perspective exercises on the of a Western woman en déshabillée (Fig. 27) adopted a same double page (Fig. 26) – also provided Hokusai’s public voyeuristic intimacy that echoed earlier bijin-ga of women with rare insights into things from worlds beyond Japan. bathing or dressing. While the casual detail of Sadahide’s Conversely, the manga provided a source of knowledge about view suggested some familiarity with his Western subject, he Japan for early European viewers following the publication was also drawing on a long ukiyo-e tradition of representing of reproductions in Philipp Franz von Siebold’s (1796–1866 women bathing or at their toilette, and on pictorial plays Nippon: Archiv zur Beschreibung von Japan (1831).74 Their on woman-and-mirror themes established by Utamaro. diversity, graphic virtuosity and rich humour ensure their Conversely, in most of these illustrations Sadahide adopts popularity among viewers of all nations today. Western representational conventions of chiaroscuro, fine Sadahide was also greatly interested in the West. Following linear hatching and linear perspective.78 the second American visit to Japan in 1854, Western visitors had an increasing impact on Japanese habits – introducing Curios: a dragon, a turtle horse racing (1862), for example, and rugby (1866). The first three volumes of Sadahide’s Yokohama kaikö kembun- and a rat shi (Observations on the opening of the Port of Yokohama) The collection’s printed works are complemented by three were published in 1862 and the second three in 1865.75 unusual painted compositions, two of which are sumi-e, Things seen and heard at the Yokahama open port is a kind of ‘monochrome ink drawings’. One of the trio is a whimsical guidebook detailing the commercial engagements, domestic arrangement of a dragon, phoenix and flowering tendrils. habits and technology of the foreign residents at Yokohama. Its seal, that of the important artist Kanö Tsunenobu Rising sun at Te Papa: the Heriot collection of Japanese art 73

(1636–1713), bears witness to the maintenance of Kanö Daikoku would have been displayed in tokonoma ‘on rat school practice throughout the Edo period. The second, by days, months, or years of the calendrical cycles used during the little-known Kagoshima literati painter Taira Natasaka the Tokugawa period’.80 Our sitter looks remarkably stately, (1780–1850), depicts a minogame, or legendary tortoise, affluent and well fed. His striped tatami mat makes a happy symbolic of long life.79 The third work, by an anonymous contrast to Kuniyoshi’s Lord En’ya, seated on his white cotton artist, seen in Fig. 28, is more mysterious. It is a small quilt and performing an infinitely grimmer task (Fig. 7). Yamato-e painting on silk, apparently of a wealthy aristocrat, perhaps a daimyö or ‘lord’, but with the head of a rat. Conclusion Whether it refers to a specific figure cannot be determined, but its design was unlikely intended to be insulting. This is Curiously, perhaps, the Japanese print has held a recurrent a sophisticated work, carefully crafted, not a frivolous object appeal for New Zealand collectors. The earliest collections or printed ephemera, and the figure itself adopts the formal of printed volumes date from the nineteenth century; dress and conventional pose of aristocratic portrait sculptures ukiyo-e generated immense enthusiasm through the first and paintings. Representations of anthropomorphic four decades of the twentieth century, and they experienced figures had a long, respectable, history in Japan, featuring a re-emergence in the aftermath of the Second World War. in early emakimono hand scrolls like the spirited, whimsical The Heriot collection signals a much more recent, and very animal tales of the Chöjü-giga (attributed to Toba Söjöō substantial, revival of this interest. The selection of these [Kukuyü], 1053–1140). Allusive animal images also populate works – in the huge majority of instances from very limited the zodiacal cycles of the calendar. The rat, for example, was supplies available in New Zealand since the 1980s – testifies associated with Daikoku, the god of wealth. Depictions of to the collecting intelligence of Ian and Mary Heriot. In turn, the selection and acquisition of items from this remarkable local collection has transformed Te Papa’s hitherto modest holdings in the area. As Ian Heriot has indicated, the pleasures generated through the last 40 years have been, in the main, privately experienced – as indeed, it had been for the earliest owners of these small, intimate works. Now, in their new home, these compositions offer opportunities to redirect their attractions to new viewers, so that they can be experienced in new contexts and curatorial combinations, thereby reviving their popularity.

Acknowledgements

The authors would like to thank Ian and Mary Heriot for their hospitality and generous advice; Andrea Hearfield, Collections Manager, Works on Paper, and the Imaging Team at Te Papa, Wellington, New Zealand. This research was supported with the award of a University of Otago Research Grant. Notes 1 These works were all in the former National Museum collection, and were amalgamated in 1992, following the establishment of the Museum of New Zealand Te Papa Tongarewa. Most of the Hiroshige prints are from the series Fifty-three Stations of the Tökaidö Road, c. 1838. Fig. 28 A Daimyö, c. 1800–40, ink and pigment on silk, 2 Ian Heriot, interview with Mark Stocker, Katikati, 11 570 x 400 mm. Artist unknown (Te Papa 2016-0008-64). June 2016. 74 Tuhinga, Number 29 (2018)

3 See especially Richard Bullen (ed.), Pleasure and play in 24 Julia Meech-Pekarik, The world of the Meiji print: Edo Japan, Christchurch: Canterbury Museum and impressions of a new civilization, New York, NY, and College of Arts, University of Canterbury, 2009. Tokyo: Weatherhill, 1987, p. 220. 4 As quoted on the Museum of New Zealand Te Papa 25 Ibid., p. 221. Tongarewa website ‘Discover the collections’ page in 26 Helen Merritt and Nanako Yamada, Woodblock kuchi-e May 2017 (link no longer active). prints: reflections of Meiji culture, Honolulu, HI: 5 Actor Posing with a Sword, c. 1835, kakemono print University of Hawai’i Press, 2000, pp. 87–89. (Te Papa 2016-0008-31). 27 Helen Merritt and Nanako Yamada, Guide to modern 6 Heriot, interview. Japanese woodblock prints: 1900–1975, Honolulu, HI: 7 Ibid. University of Hawai’i Press, 1992, p. 90. 8 Jan-Erik Nilsson, ‘Gotheborg Discussion Board 28 See David Bell, ‘Bella King in China and Japan: moderators’, in: Gotheborg antique Chinese and travelling and collecting in the early twentieth century’, Japanese collectors’ help and info page [website], 1993– New Zealand Journal of Asian Studies 14 (2), 2012, 2014, retrieved on 17 May 2017 from http:// pp. 67–85. gotheborg.com/sponsors/moderators_new.shtml. 29 See David Bell, Chüshingura and the floating world: 9 Heriot, interview. the representation of Kanadehon Chüshingura in ukiyo-e, 10 Ibid. Richmond: Japan Library, 2001; and Aleksandra 11 Ibid. Görlich, The treasury of loyal retainers: the drama of the 12 See Arthur Davison Ficke, Chats on Japanese prints, 47 ronin, Kraków: Manggha Museum of Japanese Art Rutland, VT: Charles E. Tuttle, 1958; Richard Lane, and Technology, 2012. Images from the floating world: the Japanese print, New 30 Amy Reigle Newland and Joan B. Mirviss (eds), Printed York, NY: Putnam, 1978; and Frank Whitford, to perfection: twentieth-century Japanese prints from the Japanese prints and Western painters, London: Studio Robert O. Muller collection, Washington, DC, and Vista, 1977. Amsterdam: Smithsonian Institution and Hotei 13 Heriot, interview. Publishing, 2004, p. 86. 14 For Beardsley and Japan, see, for example, Linda 31 Ibid., p. 90. Gertner Zatlin, ‘Aubrey Beardsley’s “Japanese” 32 Ibid., p. 87. grotesques’, Victorian Literature and Culture 25(1), 33 Heriot, interview. 1997, pp. 87–108. The exhibition Aubrey Beardsley and 34 Newland and Mirviss, Printed to perfection, p. 112. Japan was held at the Museum of Modern Art, Shiga, from 6 February to 27 March 2016. 35 Amy R. Newland (ed.), Visions of Japan: Kawase Hasui’s masterpieces, Amsterdam: Hotei Publishing, 15 Allen Hockley, The prints of Isoda Koryüsai: floating world culture and its consumers in eighteenth-century Japan, 2004, p. 143. Seattle, WA, and London: University of Washington 36 Junichi Ökubo, ‘Into the modern: landscape prints of Press, 2003, p. 159. Kobayashi Kiyochika and Kawase Hasui’, in: Andreas 16 Ibid., p. 3. Marks and Sonya Rhie Quintanilla (eds), Dreams and diversions: essays on Japanese woodblock prints from the 17 Ibid., p. 152. San Diego Museum of Art, San Diego, CA: San Diego 18 Ibid., p. 149. Museum of Art, p. 203. 19 Ibid., p. 154. 37 Ibid., p. 188. 20 Jacob Pins, The Japanese pillar print: hashira-e, London: Robert G. Sawers Publishing, 1982, p. 10. 38 Kendall H. Brown, ‘A place for poetry: shin-hanga 21 Hockley, Koryüsai, p. 160. landscape in modern Japan’, in Newland (ed.), Visions of Japan, p. 11. 22 Charles Baudelaire, ‘The painter of modern life’, in: Charles Baudelaire, The painter of modern life and other 39 René C. Balcer, ‘Kawase Hasui: collecting a versatile essays (ed. and trs. Jonathan Mayne), London: Phaidon modern master’, Impressions 34, 2013, p. 114. Press, 1964, p. 13. 40 Brown, ‘A place for poetry’, p. 15. 23 Helen Merritt, Modern Japanese woodblock prints: the 41 Andreas Marks, ‘Anything that creeps and f lies: natural early years, Honolulu, HI: University of Hawai’i Press, history prints’, in Marks and Quintanilla (eds), Dreams 1990, p. 4. and diversions, p. 119. Rising sun at Te Papa: the Heriot collection of Japanese art 75

42 Ibid., p. 121. 62 Matthi Forrer, Surimono in the Rijksmuseum, 43 Newland, Printed to perfection, p. 30. Amsterdam, Leiden: Hotei Publishing, 2012, p. 308. 44 Amy R. Newland, Jan Perrée and Robert Schaap, 63 Roger Keyes, ‘Tani Seikö and his circle’, Andon 72/73, Crows, cranes and camellias: the natural world of Ohara 2002, p. 12. Koson 1877–1945. Japanese prints from the Jan Perrée 64 Ibid., p. 18. collection, Leiden: Hotei Publishing, p. 28. 65 Keyes, Art of surimono, p. 503. 45 Newland et al., Crows, cranes and camellias, p. 19. 66 Jack Hillier, Japanese prints and drawings from the Vever 46 Merritt and Yamada, Woodblock kuchi-e prints, p. 61. collection. Vol. 3, London: Sotheby Parke Bernet, 1985, 47 The authors would like to thank Dr Emerald King, p. 990. Victoria University of Wellington, and Professor 67 Heriot, interview. Takashi Shogimen, University of Otago, Dunedin, 68 Jack Hillier, The art of the Japanese book, London: for their clarifications of the narrative content of this Sotheby’s Publications, 1987, p. 929. composition. 69 Richard Lane, Hokusai: life and work, London: Barrie 48 Akira Y. Yamamoto, ‘Introduction’, in: Stephen Addiss and Jenkins, 1989, p. 304. (ed.), Japanese ghosts and demons, New York, NY: 70 Ibid., p. 305. George Braziller Inc., 1985, p. 11. 71 Hillier, Japanese book, p. 815. 49 Michiko Iwasaka and Barre Toelken, Ghosts and the Japanese: cultural experience in Japanese death legends, 72 Martin Gardner, Famous poems from bygone days, Logan, UT: Utah State University Press, 1994, p. 6. Mineola, NY: Dover Publications, 1995, p. 124. 50 Joshua S. Mostow, Pictures of the heart: the Hyakunin 73 Hillier, Japanese book, p. 815. isshu in word and image Honolulu, HI: University of 74 Ibid. Hawai’i Press, 1996, p. 158. 75 Ibid., p. 929. 51 See Helen Craig McCullough, The tale of the Heike, 76 Ibid. Stanford, CA: Stanford University Press, 1998. 77 Meech-Pekarik, Meiji print, pp. 41–50. 52 Ibid., p. 284. 78 Hillier, Japanese book, p. 926. 53 See John Stevenson, Yoshitoshi’s one hundred aspects of 79 Hockley, Koryüsai, p. 155. the moon, Leiden: Hotei Publishing, 2013, p. 41. 80 Ibid. 54 Chris Uhlenbeck and Amy R. Newland, Yoshitoshi: masterpieces from the Ed Freis collection, Leiden and References Boston, MA: Hotei Publishing, 2011, p. 116. Balcer, R.C. (2013). Kawase Hasui: collecting a versatile 55 Heriot, interview. modern master. Impressions 34: 108–121. 56 Lee Bruschke-Johnson, ‘Classical themes through the Baudelaire, C. (1964). The painter of modern life. In: The prism of the “floating-world”’, in: Charlotte van painter of modern life and other essays (ed. and trs. J. Rappard-Boon (ed.), Surimono: poetry and image in Mayne). London: Phaidon Press. Pp. 1–41. Japanese prints, Leiden and Amsterdam: Hotei Bell, D. (2001). Chüshingura and the floating world: the Publishing and Rijksmuseum, 2000, p. 150. representation of Kanadehon Chüshingura in ukiyo-e 57 Ibid., p. 151. prints. Richmond: Japan Library. 170 pp. 58 Roger Keyes, The art of surimono: privately published Bell, D. (2012). Bella King in China and Japan: travelling Japanese woodblock prints in the Chester Beattie and collecting in the early twentieth century. New Library, Dublin, London: Sotheby Publications, 1985, Zealand Journal of Asian Studies 14(2): 67–85. p. 394. Brown, K.H. (2004). A place for poetry: shin-hanga landscape in modern Japan. In: Newland, A.R. (ed.). 59 Bruschke-Johnson, ‘Classical themes’, p. 151. Visions of Japan: Kawase Hasui’s masterpieces. 60 Keyes, Art of surimono, p. 394. Amsterdam: Hotei Publishing. Pp. 9–18. 61 Joan B. Mirviss, The Frank Lloyd Wright collection of Bruschke-Johnson, L. (2000). Classical themes through the surimono, New York, NY, and Phoenix, AZ: prism of the ‘floating-world’. In: van Rappard-Boon, C. Weatherhill and Phoenix Art Museum, p. 239. (ed.) Surimono: poetry and image in Japanese prints. Leiden 76 Tuhinga, Number 29 (2018)

and Amsterdam: Hotei Publishing and Rijksmuseum. Merritt, H. and Yamada, N. (1992). Guide to modern Pp. 150 –184. Japanese woodblock prints: 1900–1975. Honolulu, HI: Bullen, R. (ed.) (2009). Pleasure and play in Edo Japan. University of Hawai’i Press. 365 pp. Christchurch: Canterbury Museum and College of Arts, Merritt, H. and Yamada, N. (2000). Woodblock kuchi-e University of Canterbury. 80 pp. prints: reflections of Meiji culture. Honolulu, HI: Ficke, A.D. (2012). Chats on Japanese prints. Rutland, VT: University of Hawai’i Press. 284 pp. Charles E. Tuttle. 456 pp. Mirviss, J.B. (1995). The Frank Lloyd Wright collection of Forrer, M. (2012). Surimono in the Rijksmuseum, surimono. New York, NY, and Phoenix, AZ: Amsterdam. Leiden: Hotei Publishing. 340 pp. Weatherhill and Phoenix Art Museum. 313 pp. Gardner, M. (1995). Famous poems from bygone days. Mostow, J.S. (1996). Pictures of the heart: the Hyakunin isshu Mineola, NY: Dover Publications. 208 pp. in word and image. Honolulu, HI: University of Hawai’i Görlich, A. (2012). The treasury of loyal retainers: the drama Press. 522 pp. of the 47 ronin. Kraków: Manggha Museum of Japanese Newland, A.R. (ed.) (2004). Visions of Japan: Kawase Hasui’s Art and Technology. 256 pp. masterpieces. Amsterdam: Hotei Publishing. 149 pp. Hillier, J. (1976). Japanese prints and drawings from the Vever Newland, A.R. and Mirviss, J.B. (eds) (2004). Printed to collection. Vol. 3. London: Sotheby Parke Bernet. 207 pp. perfection: twentieth-century Japanese prints from the Hillier, J. (1987). The art of the Japanese book. London: Robert O. Muller collection. Washington, DC, and Sotheby’s Publications. 1131 pp. Amsterdam: Smithsonian Institution and Hotei Hockley, A. (2003). The prints of Isoda Korüysai: floating Publishing. 130 pp. word culture and its consumers in eighteenth-century Japan. Newland, A., Perrée, J. and Schaap, R. (2001). Crows, cranes Seattle, WA, and London: University of Washington and camellias: the natural world of Ohara Koson 1877– Press. 313 pp. 1945. Japanese prints from the Jan Perrée collection. Iwasaka, M. and Toelken, B. (1994). Ghosts and the Leiden: Hotei Publishing. 208 pp. Japanese: cultural experience in Japanese death legends. Nilsson, J.E. (1993–2014). ‘Gotheborg Discussion Board Logan, UT: Utah State University Press. 138 pp. moderators’. In: Gotheborg antique Chinese and Japanese Keyes, R. (1985). The art of surimono: privately published collectors’ help and info page [website]. Retrieved on Japanese woodblock prints in the Chester Beattie Library, 17 May 2017 from http://gotheborg.com/sponsors/ Dublin. London: Sotheby Publications. 569 pp. moderators_new.shtml. Keyes, R. (2002). Tani Seikö and his circle. Andon 72/73: Ökubo, J. (2010). Into the modern: landscape prints of 12–24. Kobayashi Kiyochika and Kawase Hasui. In: Marks, A. Lane, R. (1978). Images from the floating world: the Japanese and Quintanilla, S.R. (eds). Dreams and diversions: essays print, including an illustrated dictionary of ukiyo-e. New on Japanese woodblock prints from the San Diego Museum York, NY: Putnam. 364 pp. of Art. San Diego, CA: San Diego Museum of Art. Lane, R. (1989). Hokusai: life and work. London: Barrie and Pp. 187–205. Jenkins. 320 pp. Pins, J. (1982). The Japanese pillar print: hashira-e. London: McCullough, H.C. (1998). The tale of the Heike. Stanford, Robert G. Sawers Publishing. 387 pp. CA: Stanford University Press. 489 pp. Stevenson, J. (2013). Yoshitoshi’s one hundred aspects of the Marks, A. (2010). Anything that creeps and flies: natural moon. Leiden: Hotei Publishing. 272 pp. history prints. In: Marks, A. and Quintanilla, S.R. (eds). Uhlenbeck, C. and Newland, A.R. (2011). Yoshitoshi: Dreams and diversions: essays on Japanese woodblock masterpieces from the Ed Freis collection. Leiden and prints from the San Diego Museum of Art. San Diego, Boston, MA: Hotei Publishing. 176 pp. CA: San Diego Museum of Art. Pp. 119–128. Whitford, F. (1977). Japanese prints and Western painters. Meech-Pekarik, J. (1987). The world of the Meiji print: London: Studio Vista. 264 pp. impressions of a new civilization. New York, NY, and Yamamoto, A.Y. (1985). Introduction. In: Addiss, S. (ed.). Tokyo: Weatherhill. 260 pp. Japanese ghosts and demons. New York, NY: George Merritt, H. (1990). Modern Japanese woodblock prints: the Braziller Inc. Pp. 9–13. early years. Honolulu, HI: University of Hawai’i Press. Zatlin, L.G. (1997). Aubrey Beardsley’s ‘Japanese’ 324 pp. grotesques. Victorian Literature and Culture 25(1): 87–108. Tuhinga 29: 77–89 Copyright © Museum of New Zealand Te Papa Tongarewa (2018) 77

Characterisation of 15 sources of Japanese obsidian: PIXE–PIGME analysis, and identification of archaeological specimens Foss Leach,* Shizuo Oda** and Roger Bird***

* Research Associate, Museum of New Zealand Te Papa Tongarewa, PO Box 467, Wellington, New Zealand ([email protected])

** University Museum the University of Tokyo, Japan

*** Formerly, Australian Atomic Energy Commission Research Establishment, Private Mailbag, Sutherland, NSW 2232, Australia

ABSTRACT: Samples of obsidian from 15 different volcanic sources in Kyushu, Honshu and Hokkaido were subjected to analysis using the PIXE–PIGME method. Eighteen elements were resolved for each source, and a series of statistical techniques was used to assess the distinctiveness of the sources. It was found that the sources could be separated into three groups, each with sub-sources that are virtually indistinguishable from each other. The first group of four sources originates in the Shinshu area, the second group of two sources is from the Izu Islands, and the third group contains two sources on Kyushu. Each of these groups is easily distinguished from the others. Finally, nine obsidian artefacts from three archaeological sites were analysed and compared with these sources. Eight of these could be allocated to a source, while one proved to be from an unknown source.

KEYWORDS: Obsidian, Japan, PIXE, PIGME, archaeology, source characterisation.

Introduction The main purpose of the study was to examine the range of variation of element concentrations from one source to another to see how clearly they can be distinguished. Obsidian is a form of volcanic glass that can readily be flaked Characterisation of obsidian sources is the first step before into tools with extremely sharp edges. It is found widely attempting to identify the origin of archaeological specimens. across the Pacific, as well as elsewhere in the world, and As Kuzmin has pointed out (2012: 1), it may be impossible was used by prehistoric people as cutting implements. There to distinguish between some primary sources of obsidian are thought to be at least 80 sources of obsidian of varying unless a wide range of elements is determined, using such quality throughout the islands of Japan. In this study we methods as neutron activation analysis (NAA) and laser focused on 15 sources: eight from Honshu, two from the ablation inductively coupled plasma mass spectrometry Izu Islands, three from Kyushu and two from Hokkaido. (LA-ICP-MS). However, it is sensible to use a broad-brush This represents a suitable sample of the range of geochemical strategy over the range of variation involved first, to narrow variation that might be encountered when trying to identify down sources that are difficult to distinguish other than by archaeological specimens. using more expensive techniques. 78 Tuhinga, Number 29 (2018)

Fig. 1 Map of Japan showing the location of the obsidian sources analysed. Sources of Japanese Obsidian 79

The volcanic sources sampled Izu Islands 9 Sanuka, Közu Island (6). The location of each obsidian source is shown in Figure 1. 10 Sawajiri Bay, Közu Island (5). They are enumerated below. The numeral in brackets is the number of specimens analysed from each source (total=82). Additional specimens from Nagahama and Onbase in the Izu Islands are very poor quality, crystalline and have many Honshu inclusions, and were not analysed. 1 Hoshigatö, Shinshu Area (5). Kyushu 2 Azeria, Shinshu Area (4). 11 Koshidake, Saga Prefecture (6). 3 Higashi Mochiya A, Shinshu Area (6). 12 Yodohime, Sasebo City, 4 Higashi Mochiya B, Shinshu Area (6). Nagasaki Prefecture (6). 5 Kobukazawa, Shinshu Area (5). 13 Muta, Nagasaki Prefecture (5). 6 Omegura, Shinshu Area (5). Hokkaido 7 Mugikusa Toge, Shinshu Area (7). 14 Tokachi (6). 8 Kamitaga, Hakone Region (4). 15 Oketo (6).

The PIXE–PIGME analytical method

The PIXE (proton induced X-ray emission) and PIGME (proton induced γ-ray emission) analytical techniques have been outlined in previous publications: the PIXE method by Duerden et al. (1979) and the PIGME method by Bird et al. (1978). The use of the joint PIXE–PIGME technique for elemental concentration measurements has been described by Duerden et al. (1980). The experimental arrangement for obsidian specimens ready to be analysed in this study is shown in Fig. 2. The γ-ray and X-ray spectra collected were automatically recorded. Batches of spectra were later reduced interactively on a computer with the aid of graphics. With the use of suitable standards, absolute concentrations were determined. A beam of 100–200 µ A of 2.5 MeV protons was used to irradiate each obsidian sample and µ-ray and X-ray detectors measured the radiation emitted during irradiation. A typical running time was five minutes per sample. Samples were pre-mounted onto target holders Fig. 2 The sample chamber for the PIXE–PIGME analysis and these, in turn, were located by pegs on a target stick of obsidian objects. (Fig. 3), where they were held in place by magnetic 80 Tuhinga, Number 29 (2018)

Fig. 3 Up to 60 obsidian samples at a time were mounted on a special target stick, which was then inserted into the sample chamber for analysis. tape. The smoothest face of the sample was selected to be objects selected had to fit into the side tubes of the target irradiated. changer. A maximum of 60 samples can be loaded onto the The samples were pre-cleaned and the entire loaded target stick and, with automatic sequencing, up to 200 or target stick washed by ultrasonic agitation in nanograde 300 samples can be studied per day. benzene. Although surface irregularities were minimised The detectors were placed at 135 degrees to the by cleaning, the objects were often irregular and sometimes direction of the proton beam. The 120 cm3 Ge(Li) detector had rough faces. The effect of roughness has been monitored used for detection of the γ-rays had a resolution of 2.3 keV statistically and has generally been found to have minimal at 1.33 MeV and an efficiency of 1.9 x 104 peak counts per effect (Bird et al. 1978). The only limitation was that the γ-ray for a source of the energy placed at 25 cm. The 12

Fig. 4 PIGME spectrum of an obsidian sample. Sources of Japanese Obsidian 81

mm2 Si(Li) detector with 4 mm depletion depth for X-ray The detected X-ray spectrum is dominated by K detection was mounted via a sliding seal on a chamber port. X-rays1 from the major elements (typically Al, Si, K, Ca, Its resolution was 140 eV at the Mn Kσ 5.9 keV line. The Ti, Fe), whose energies are in the range 1.5–7.0 keV. K efficiency of the Si(Li) detector with and without filters X-rays from Zn, Pb, Sr, Y, Zr and Nb elements that appear was measured and the efficiency functions determined. to be significant for sourcing obsidian are many orders of A heated graphite filament was used to discharge any magnitude lower than the previously mentioned elements charging that could result from the proton irradiation and cannot be measured simultaneously without the use of a of an insulated target, which would otherwise affect the special filter. A 1 mm graphite filter was used, with a 75 µm X-ray background. The target chamber was used as a hole laser-drilled at the centre. This allows low-energy Faraday cup for collection of target current, a technique X-rays (<6 keV) to be observed with approximately 0.1% that is not completely satisfactory and that can certainly be of their normal intensity at the same time as the higher- affected by the electron flood current. A shift in absolute energy X-rays, which are little affected by the filter. The concentration values was seen in a limited number of resulting spectrum is shown in Fig. 5. repeated measurements. The detected γ-ray and X-ray spectra were analysed in A typical γ-ray spectrum from an obsidian sample is part by simple data reduction during the experiment. For shown in Fig. 4. It can be seen that γ-rays from Na, Al and the γ-ray spectrum, counts of the Na 440 keV and 1630 F dominate the spectrum and, although in some particular keV peaks, the Al 643 keV and 1013 keV peaks, and the cases other γ-rays may be measured with reasonable three higher-energy peaks of F were derived and stored precision, only the X-rays from the three elements (Na, Al, on file. The X-ray data were reduced to 24 summations F) are measured on a routine basis for obsidian. of counts between fixed energy equivalent ‘windows’ set

Fig. 5 (left) PIXE spectrum at 1–10 keV of a piece of obsidian. The dashed line is the experimental data, and the solid line is the analysed best fit to the data and the background component. 82 Tuhinga, Number 29 (2018)

about prominent peaks in the spectrum. These figures were Gamma-ray measurements also stored on file, together with run information: sample Na ±20%; Al, F ±3% number, run number, current, elapsed time for the counting X-ray measurements period, etc. The complete X-ray spectrum was then stored Rb, Mn, Fe ±5%; K ±6%; Si, Ca ±7%; Sr ±8%; Zr ±9%; for a more complete analysis offline. This analytical procedure has been described in detail Ga ±12%; Ti ±20%; As, Pb, Br, Zn ±30%; Y and Cu by Clayton (1982a,b), and essentially consists of fitting a usually low concentration, so poor precision ±~50% predetermined background shape and Gaussians representing characteristic X-ray peaks to the spectra. The relative Element results for source samples intensity and the energies of the characteristic X-rays for A total of 82 obsidian samples from the 15 sources were a particular element are fixed, but the overall magnitude analysed. Each sample was allocated a catalogue number of the peaks, and of the background, are scaled to best fit beginning with two letters (JD, JE, etc.), so that individual the data. Once again, the resulting analysed elemental data items can be referred to. The Japan samples were part of were stored on file. An example of the fit to a typical X-ray a larger run that included an additional 567 samples from spectrum is shown in Fig. 5. It can be seen that the peak/ other parts of the Pacific region (see Fig. 6 and Duerdenet background ratio is much higher for some elements than al. 1987). others. The precision of measurement within the spectra therefore varies from element to element, but can also The obsidian samples from two additional sources from change from spectrum to spectrum because of the relative Nagahama and Onbase in the Izu Islands were of such a low element concentrations. quality that no valid measurement could be made with our A full calculation was made for the number of X-rays analysis technique. The element concentrations for samples detected per 100 µC per 100 µg/g of a trace element within from a source were generally in agreement with one another an obsidian sample. These yield values were then used to within the errors of the measurement technique. There convert all analysed data to elemental concentration either were, however, three anomalous results: in percentage by weight or µg/g. 1 Higashi Mochiya A, sample JD405. Most element Precision of results concentrations are in agreement with other JD samples, but Ca is 1.12%, whereas for other samples the average A standard overall error of ±5% is present for all Ca is 0.46±0.02, and Sr is 108 g/g, compared with concentrations listed. This primarily results from inaccuracy 17±7 g/g for other samples from the same source. of current measurements. However, within an analysis, an 2 Higashi Mochiya B, sample JE503. Most elements are improved accuracy can be expected in element ratios. No in agreement with other JE samples, but Zr is 216 g/g, repeat measurements were made, but from measurements compared with 83±4 g/g for other samples. made on different rocks from the same source we can make 3 Higashi Mochiya B, sample JE508. The values for Ti, some generalised comments about measurement precision. Fe and Sr for this sample appear to be very different to others from the same source. The final precision is therefore suggested to be:

Table 1 Mean element concentrations for the 15 Japanese obsidian sources sampled (Fig. 1). Note that the mean is listed as zero for source 8 as no value was determined. Major elements (%) Minor elements (µg/g) Source Na Al Si K Ca Fe F Ti Mn Zn Ga As Pb Rb Sr Y Zr Nb 1 2.55 6.14 37.72 3.86 0.41 0.58 663.1 679.0 810.1 26.4 13.3 14.5 4.7 149.0 42.0 18.8 75.9 3.7 2 2.68 6.39 38.78 3.97 0.41 0.64 1140.7 424.4 1162.9 28.0 24.0 15.8 4.4 315.6 6.6 24.0 94.5 7.1 3 2.56 6.23 35.40 3.59 0.57 0.67 1021.1 673.8 1023.5 25.4 19.0 16.0 1.4 277.5 17.0 23.1 80.3 9.3 4 2.55 6.27 36.62 3.66 0.53 0.79 1046.2 458.5 1064.4 21.1 18.1 19.0 2.1 279.1 13.7 27.2 104.9 5.6 5 2.67 6.36 40.36 4.00 0.47 0.67 1111.7 617.3 1155.3 26.3 21.9 24.6 4.0 305.3 10.4 22.4 83.4 5.9 6 2.53 6.24 36.74 3.70 0.58 0.84 556.9 1235.0 765.1 21.8 10.4 12.3 5.1 163.8 82.8 11.6 127.7 2.5 7 2.64 6.36 36.24 3.17 0.64 0.88 596.2 1412.1 573.5 17.9 9.4 17.4 6.0 101.0 136.8 16.7 114.6 3.7 8 3.11 6.30 36.89 1.37 1.29 1.45 379.6 1288.9 579.0 16.7 8.5 7.3 3.1 31.9 142.6 25.9 144.4 0.0 9 2.92 6.47 41.39 3.12 0.63 0.68 288.8 1008.1 867.9 17.1 1.9 6.7 0.6 75.4 67.7 20.9 58.7 5.5 10 2.88 6.25 38.64 2.70 0.65 0.72 273.9 866.2 859.1 17.7 3.7 4.8 1.0 69.5 76.9 16.4 71.6 2.9 11 2.62 6.64 39.76 4.02 0.54 1.04 557.7 345.2 524.5 25.5 13.8 21.0 2.7 204.2 43.8 8.9 61.6 14.7 12 2.83 6.88 36.29 3.47 0.76 1.33 386.0 659.6 537.5 23.3 10.6 16.8 1.7 144.3 106.1 9.2 117.2 11.1 13 2.57 6.61 37.77 3.90 0.56 1.00 557.2 443.5 481.5 22.3 12.6 20.5 2.7 201.6 46.9 10.8 83.4 11.6 14 2.47 6.20 37.69 3.31 0.70 0.97 280.0 892.6 430.9 18.1 6.7 15.7 1.6 151.7 64.9 14.1 94.7 1.4 15 2.71 6.47 38.94 3.48 0.60 0.98 428.5 545.2 519.4 21.0 7.1 18.2 1.7 160.0 56.6 21.5 76.0 1.5 Sources of Japanese Obsidian 83

Table 2 Standard deviations of element concentrations for the 15 Japanese obsidian sources sampled (Fig. 1). Note that the standard deviation is listed as zero for sources 8 (no value determined) and 9 (only one value determined).

Major elements (%) Minor elements (µg/g) Source Na Al Si K Ca Fe F Ti Mn Zn Ga As Pb Rb Sr Y Zr Nb 1 0.12 0.26 2.90 0.21 0.03 0.01 26.6 53.9 38.1 0.4 0.6 3.3 0.02 7.3 3.5 7.0 27.0 3.2 2 0.05 0.18 1.58 0.05 0.01 0.02 19.5 63.6 67.3 6.2 5.3 6.0 2.7 5.6 3.1 8.5 11.0 3.7 3 0.05 0.30 3.11 0.39 0.27 0.06 40.1 157.4 54.3 2.8 2.6 4.1 0.6 13.3 6.2 7.5 3.2 5.6 4 0.03 0.19 2.21 0.22 0.17 0.35 10.2 132.2 24.2 5.4 3.2 6.4 2.0 17.2 1.5 4.2 54.8 2.5 5 0.05 0.13 4.03 0.36 0.07 0.05 22.1 24.5 81.2 10.0 4.3 12.8 3.4 11.8 4.5 4.9 3.7 3.6 6 0.03 0.17 2.19 0.24 0.02 0.02 21.3 145.3 15.3 0.8 2.1 2.4 1.2 8.6 12.0 5.6 13.3 2.0 7 0.06 0.25 4.05 0.45 0.06 0.06 18.8 222.0 16.9 4.1 2.0 5.8 1.0 4.5 12.7 4.1 12.3 2.3 8 0.06 0.25 3.61 0.07 0.17 0.11 35.9 105.6 35.8 0.7 1.7 1.4 0.3 4.2 11.5 6.1 10.2 0.0 9 0.15 0.21 3.92 0.21 0.06 0.03 10.5 95.0 63.9 1.8 0.7 2.1 0.2 7.2 10.2 0.9 7.7 0.0 10 0.01 0.04 2.21 0.10 0.05 0.04 8.5 95.0 45.7 0.7 0.4 2.0 0.9 6.1 3.1 2.1 5.0 0.9 11 0.04 0.23 4.09 0.41 0.06 0.11 19.6 139.6 37.4 2.8 2.0 3.5 0.7 7.8 2.0 3.4 7.1 2.1 12 0.02 0.17 3.24 0.24 0.06 0.03 7.7 138.6 16.1 1.4 2.3 3.1 0.5 3.4 3.3 3.0 12.0 2.5 13 0.05 0.09 2.99 0.17 0.03 0.06 6.9 192.6 47.2 5.9 4.4 6.6 1.0 9.9 4.2 3.5 21.9 7.2 14 0.02 0.12 2.75 0.22 0.05 0.03 2.2 149.7 35.2 0.6 1.1 1.7 0.7 3.4 4.7 3.5 6.7 1.8 15 0.02 0.06 2.29 0.21 0.02 0.02 6.8 147.2 22.0 1.6 2.5 3.8 0.9 3.0 1.7 4.6 3.4 1.2

Comparison with other compared in Table 3. There is good agreement between the two analyses, although the Sr values determined by published research NAA seem rather low compared with those by PIXE– Kuzmin et al. (2013: 361) provide useful comparative PIGME in the case of the Shirataki series. However, Sr is data from their NAA analysis of samples of obsidian from not an easy element to resolve in an obsidian matrix using four sources on Hokkaido, two at Oketo, and two from NAA, and results are not very reliable (Leach 1996: 88; the Shirataki region. Our Tokachi source material is also Leach & Warren 1981). The peak is very close to three from the Shirataki field, but about 3.5 km southeast of the other isotopes (131Ba, 65Zn, 85Sr and 147Nd, situated at 496 Kuzmin et al. sampling areas. There are nine elements in keV, 511 keV, 514 keV and 531 keV, respectively), and common between the NAA study by Kuzmin et al. and appears as only a small peak above background after 30 our own PIXE-PIGME analysis; the two sets of results are days, when long-lived isotopes are counted.

Table 3 Comparison of two studies of Hokkaido obsidians. A = PIXE–PIGME results, this study (mean values); B = NAA results from Kuzmin et al. 2013: 361 (mean values).

Na Al K Mn Fe Zn Rb Sr Zr ±20% ±3% ±6% ±5% ±5% ±30% ±5% ±8% ±9% Tokachi/ Shirataki A 2.47 6.20 3.31 431 0.972 18.1 152 66 95 B 2.87 6.80 3.82 384 0.788 37.5 150 28 90 B 2.94 6.85 3.85 451 0.743 35.6 174 <15 87 Oketo A 2.71 6.47 3.48 519 0.982 21.0 160 57 76 B 2.77 6.55 3.46 321 0.732 27.0 135 74 116 B 3.13 6.60 2.84 371 0.900 37.0 98 89 129 84 Tuhinga, Number 29 (2018)

The geochemical character The so-called mildly peralkaline obsidians (comendites) have an agpaitic index of between 1.00 and 2.50. Finally, of the obsidian it should be noted that the agpaitic index has also been used (by Smith et al. 1977) to distinguish between oceanic Volcanic glasses have been classified in a number of ways. obsidians from diverging or tensional (extensional) plate For example, the weight percentage of silica is used to boundaries, and circum-oceanic obsidians from converging distinguish between three types (after Macdonald 1974: or compressional (subducting) plate boundaries: 500 ff.): 6 Oceanic obsidians >1.00 agpaitic index 7 Circum-oceanic obsidians <1.00 agpaitic index 1 Rhyolitic obsidian >66 wt% SiO2 2 Dacitic obsidian <66 >52 wt% SiO 2 The data in Table 1 were used to calculate these molecular 3 Tachylite (basaltic glass) <52 wt% SiO 2 ratios. The results are plotted in Fig. 6, together with other data from Pacific Island obsidians (Leach et al. 1986). In addition, the agpaitic index – the molecular ratio of It can be seen that, with the exception of three marginal (Na O + K O)/A1 O – has proved to be a useful means 2 2 2 3 pieces, the Japanese obsidian is clearly classed as rhyolitic of distinguishing between: and alkaline, and belongs to the circum-oceanic group. The three items that are slightly to the left of the dacitic/ 4 Peralkaline obsidian >1.00 agpaitic index rhyolitic boundary are one piece from Higashi Mochiya A, 5 Alkaline or calc-alkaline <1.00 agpaitic index one from Mugisaka and one from Yodohime.

Fig. 6 Classification of Japanese and other Pacific obsidian samples. The obsidian samples from Japanese sources are shown in red, while the blue dots indicate other obsidian samples from elsewhere in the Pacific region. Sources of Japanese Obsidian 85

Discrimination between Such correlations vary from one source of obsidian to another, thus negating one of the important assumptions the 15 sources involved in many statistical techniques commonly used The different elements resolved through any method when assessing similarity and dissimilarity between groups of analysis do not have equal power to discriminate of multivariate data, that is a uniform variance–covariance between different sources. One obvious reason is that the matrix. Linear discriminant function analysis is therefore concentration of any one element may vary so much from invalidated. This problem was addressed by Leach and one specimen to another that the different sources appear Manly (1982), and a minimum Mahalanobis distance similar. A less obvious reason is that the concentration function was developed to overcome such problems. This of some elements can be highly correlated with others, resulted in a nest of three algorithms named Popper’s razor so the additional elements do not add new information 1, 2 and 3.2 This approach was used here to discriminate when trying to discriminate between sources. There are between the 15 obsidian sources from Japan. The results several such cases of high correlation among these sources are given in Table 4. of obsidian from Japan. These are listed below (r >0.6): Table 4 show that there is considerable overlap between sources 2, 3, 4 and 5 (Azeria, Higashi Mochiya A, Higashi K/Si 0.97 Mochiya B and Kobukazawa, all from Shinshu), which Mn/K 0.92 clearly form a close-knit group, with considerably higher Al/Na 0.90 values for F, Mn and Rb than other sources. We should Mn/Si 0.90 remember that F and Rb are highly correlated (r = 0.85), Ca/Al 0.85 so it is really only two elements that are higher than in these Fe/Ca 0.85 sources than in others. Rb/F 0.85 There is also about 25% overlap between sources 9 and Fe/Al 0.83 10 (Sanuka and Sawajiri Bay on Közu Island in the Izu Ca/Na 0.73 group). These two sources have fairly distinctive values of Mn and Rb, and low values of Ga and Pb. Fe/Na 0.67

Table 4 Overlaps between one source and another. The estimated number of individuals that would be closer to the centroid of another source than to that of the correct source is shown below, for 1000 individuals. This is calculated from the probability of overlap in an 18-dimensional hyperspace from Mahalanobis D2.

Source 1 2 3 4 5 6 7 8 9 10 11 12 13 14 2 0 3 0 0 4 0 8 401 5 0 244 34 165 6 1 0 0 0 0 7 0 0 0 0 0 2 8 0 0 0 0 0 0 3 9 0 0 0 0 0 0 0 0 10 0 0 0 0 0 0 0 0 243 11 8 0 0 0 0 0 0 0 0 0 12 0 0 0 0 0 0 19 0 0 0 0 13 9 0 0 0 0 0 0 0 0 0 840 0 14 0 0 0 0 0 0 4 0 0 0 0 3 0 15 1 0 0 0 0 0 0 0 0 0 0 2 5 36 86 Tuhinga, Number 29 (2018)

Finally, sources 11 and 13 (Koshidake and Muta Identification of the volcanic on Kyushu) are clearly extremely similar, with about 85% overlap. They have quite distinctive Rb and Sr source of obsidian artefacts concentrations. A small sample of nine obsidian artefacts was selected to Apart from these three covarying groups, which test whether their sources could be determined using the are clearly quite similar geochemically, there is little PIXE–PIGME data. The artefacts were from three different multivariate overlap between other sources, or indeed archaeological sites, all dating to the Middle Jömon period between any one of the three groups and another. This is a (c. 5000 yrs bp), based on associated pottery finds: very satisfying result. Canonical analysis was performed on the element data A Three specimens from the Togari Ishi Museum, of the 15 sources. The first eigenvector extracted 86% of Parking Lot, Nagano Prefecture. Catalogue numbers the total variance, the first two eigenvectors 95% and the JA101, JA102, JA103. first three 98%. Clearly, the multivariate relationships can B Two specimens from the Maehara Site, Koganei city, be depicted quite satisfactorily in three dimensions. This is Tokyo, Middle Jömon, fourth pit dwelling. Catalogue shown in Fig. 7. numbers JG702, JG703.

C Four specimens from the Togari Ishi Site surface collection, Nagano Prefecture. Catalogue numbers JK101, JK102, JK103, JK104.

Fig. 7 Canonical analysis showing the relationship between different sources of obsidian from Japan. Three eigenvectors account for 98% of the total variance. There are three groups of sources that share significant similarity: sources 2, 3, 4 and 5; sources 9 and 10; and sources 11 and 13. Sources of Japanese Obsidian 87

Table 5 Element concentrations for nine obsidian artefacts from Japan.

Major elements (%) Minor elements (µg/g) Artefact Na Al Si K Ca Fe F Ti Mn Zn Ga As Pb Rb Sr Y Zr Nb

JA101 2.34 5.92 33.59 3.31 0.50 0.77 615.0 1214.1 518.2 15.0 0.0 0.0 0.0 98.2 101.5 20.6 106.9 13.2

JA102 2.49 5.81 35.63 3.18 0.55 0.79 500.9 1202.8 504.1 19.7 0.0 0.0 0.0 88.6 101.5 24.8 117.4 8.3

JA103 2.76 6.59 39.54 3.45 0.65 0.88 571.9 1558.7 600.0 11.3 10.7 8.5 7.6 116.0 145.5 17.9 128.3 13.2

JG702 2.61 6.17 39.26 3.89 0.47 0.67 1084.1 747.4 1168.7 15.7 19.3 14.3 2.1 300.3 12.2 36.7 109.0 19.8

JG703 2.61 6.06 38.98 4.01 0.43 0.56 656.1 840.6 787.4 21.6 6.9 16.3 0.0 157.2 38.0 7.3 70.3 10.4

JK101 2.60 6.70 34.17 3.37 0.97 0.80 1016.0 883.3 1035.7 16.0 14.0 17.4 3.4 267.6 95.4 32.4 85.2 25.4

JK102 2.61 6.35 39.87 3.96 0.39 0.60 673.0 1053.2 805.7 18.8 7.1 13.3 3.4 157.1 46.0 19.4 93.9 13.5

JK103 2.59 6.11 36.60 3.56 0.38 0.58 669.0 787.2 757.3 18.2 11.9 6.1 2.6 145.1 45.1 18.2 71.9 5.9

JK104 2.57 6.20 40.30 3.95 0.47 0.56 641.7 802.6 807.7 23.4 7.8 15.8 0.5 157.8 46.8 15.9 74.5 5.3

The Mahalanobis D 2 distances were calculated between the at Hoshigatō, and specimen JK104 from the same surface artefact and the centroid of each of the 15 sources, divided collection probably came from this source, but is possibly by the effective number of dimensions in each case. If an significantly different. Maehara is 118 km, and Togari Ishi artefact actually belongs to a particular source, the distance 64 km, from Hoshigatö. should be approximately 1.0. Specimen JG702 from Maehara is located in the The significance of theD 2 distance is also calculated. The overlapping regions of two sources 4 and 5, Higashi Mochiya results are given in Table 6. B and Kobukazawa, and is not significantly different from The first three specimens – JA101, JA102, JA103, either. The Maehara site is 64 km from both these sources. from the Togari Ishi Museum, Parking Lot site – are not Specimen JK101 cannot be allocated to any of the significantly different from source number 7 at Mugikusa known sources in this study Toge, 64 km away. It is interesting that artefact JG702 appears to be so similar Specimens JG703 from the Maehara site, and JK102 to two of the sources, with little to separate the alternatives and JK103 from the surface collection at the Togari Ishi (2.1 and 2.2 Mahalanobis D2 distances, respectively). site are not significantly different from source number 1 In another context, when trying to locate the origin of

Table 6 Identification of the source of each artefact. Conventions for significance are: not significant (NS) =≥ 5%; possibly significant (PS) = < 5%≥ 1%; significant (S) = < 1%≥ 1%; highly significant (HS) = < 1%.

Artefact Source D2 Significance Source D2 Significance JA101 7 2.2 10 NS JA102 7 2.1 10 NS JA103 7 2.5 5 NS JG702 4 2.1 10 NS 5 2.2 10 NS JG703 1 2.7 5 NS JK101 1 7.4 0 HS JK102 1 2.0 10 NS JK103 1 2.1 10 NS JK104 1 3.4 1 PS 88 Tuhinga, Number 29 (2018) a floating piece of obsidian (Leach et al. 2016), it was Conclusions useful to consider the proportional difference between the element signature of an item and possible sources. Thus, the Japan has a large number of obsidian sources. The 15 sources concentration of an element in the item being considered studied here probably represent only about one-fifth of those = C1, the mean concentration of a sample from the source known to exist. It is not known how many of the total will being considered = C2, and the absolute difference (∆1) = prove to be geochemically distinguishable from one another, abs(C2–C1). The proportional difference is then Δ∆1/C1. but this study suggests that some sources will prove difficult to Such a method standardises differences, so that an element separate using element data. Many of the 15 sources studied with, say, a concentration of 12µg/g will have the same can be clearly differentiated using the 18 elements analysed weight as another element with a concentration of 2000 by PIXE–PIGME, but three groups of sources are extremely µg/g. The means and standard deviations of the source similar in composition. Fortunately, the sources in each of samples are also calculated. The standard deviations are then these groups are geographically close to one another, and for standardised as a proportion of the mean for each element. the purposes of identifying the movement of raw material This permits us to visualise the variability of any one element used to make obsidian artefacts found in archaeological sites, using a standard point of reference. This is attempted in Fig. such groupings are not significant. Other than these three 8 using the data from artefact JG702, plotting it against the groups, there is practically no multivariate overlap between standardised data for the two alternative sources: Higashi the sources studied, so the task of differentiating the sources Mochiya B (source 4) and Kobukazawa (source 5). This using element characteristics has been very successful. plot emphasises two points: these sources are very similar to Finally, the small number of archaeological artefacts each other, and the composition of the artefact is so similar that were analysed as a test case showed that their source to both that distinguishing between the two alternatives is could be determined with acceptable accuracy. In one case, not possible. the artefact proved to be significantly different to any of Fig. 8 (below) Comparison of artefact JG702 with two the 15 sources studied. This underlines the need to gather alternative sources, Higashi Mochiya B and Kobukazawa, samples from many more volcanic sources in Japan and to using standardised data. characterise these in the future. Sources of Japanese Obsidian 89

Acknowledgements Duerden, P., Clayton, E., Bird, J.R., Ambrose, W. and Leach, B.F. (1987). Obsidian composition catalogue. Unfortunately, one of the authors, Roger Bird, was Pp. 232–38. In: Ambrose, W.R. and Mummery, J.M.J. deceased by the time this research reached publication stage. (eds). Archaeometry: further Australasian studies. We would like to acknowledge the enthusiasm and support Canberra: Department of Prehistory, Australian National Dr Bird gave to the development of the obsidian-sourcing University. 350 pp. project over a long period, and the development of PIXE– Kuzmin, Y.V. (2012). Long-distance obsidian transport PIGME as a means of assisting archaeologists in tracing the in prehistoric northeast Asia. Bulletin of the Indo-Pacific movement of raw materials in the Pacific region. Prehistory Association 32: 1–5. We would also like to thank the staff of Lucas Heights for Kuzmin, Y.V., Glascock, M.D., and Izuho, M. (2013). the use of their 3-MeV accelerator. The samples of obsidian The geochemistry of the major sources of archaeological studied in this paper were collected by Shizuo Oda and were obsidian on Hokkaido Island (Japan): Shirataki and given to the senior author for analysis during an expedition Oketo. Archaeometry 55(3): 355–369. to the Ogasawara Islands, southern Japan, in September Leach, B.F. (1996). New Zealand and oceanic obsidians: an 1989. Although the analysis was carried out shortly after archaeological perspective using neutron activation that, it has taken until now to write up the results. analysis. Journal of the Royal Society of New Zealand 26(1): 79–105. Notes Leach, B.F. and Manly, B. (1982). Minimum Mahalanobis 1 K X-rays are X-rays that arise from electron transitions distance functions and lithic source characterization by in the innermost electron shell, known as the K shell. multi element analysis. New Zealand Journal of Archaeology 4: 77–109. 2 These programs, together with test input and output Leach, B.F. and Warren, S.E. (1981). Neutron activation files, may be downloaded from http://waihinga.ac.nz/ analysis of New Zealand and oceanic obsidians: towards foss/university-of-ngakuta/downloads/Poppers- a simple screening technique. Pp. 151–166 In: Leach, Razor.zip. B.F. and Davidson, J.M. (eds). Archaeological studies of References Pacific stone resources. British Archaeological Reports Bird, J.R., Russell, L.B., Scott, M.D. and Ambrose, W.R. S104. Oxford: British Archaeological Reports. 237 pp. (1978). Obsidian characterization with elemental Leach, B.F., Anderson, A.J., Sutton, D.G., Bird, J.R., analysis by proton-induced gamma-ray emission. Duerden, P. and Clayton, E. (1986). The origin of Analytical Chemistry 50: 2082–2084. prehistoric obsidian artefacts from the Chatham and Clayton, E. (1982a). Sample characterization by proton Kermadec islands. New Zealand Journal of Archaeology induced X-ray emission analysis. Application of Surface 8: 143–170. Science 13: 136–158. Leach, B.F., Campbell, H., Eby, N., Holt, K., Regelous, Clayton, E. (1982b). Some statistical techniques for M., Richards, R., and Weaver, S. (2016). Obsidian provenancing artefact material: a user’s view. Pp. 90–99. floater washed up on a beach in the Chatham Islands: In: Ambrose, W.R. and Duerden, P. (eds). Archaeometry: geochemical composition and comparison with other an Australasian perspective. Canberra: Department of volcanic glasses. Tuhinga: Records of the Museum of New Prehistory, Australian National University. 391 pp. Zealand Te Papa Tongarewa 27: 21–49 Duerden, P., Cohen, D.D., Clayton, E., Bird, J.R. and Macdonald, R. (1974). Nomenclature and petrochemistry Leach, B.F. (1979). Elemental analysis of thick obsidian of the peralkaline oversaturated extrusive rocks. Bulletin samples by proton induced X ray emission spectrometry. Volcanologique 38(3): 498–516. Analytical Chemistry 51(14): 2350–2354. Smith, I.E.M., Ward, G.K. and Ambrose, W.R. (1977). Duerden, P., Bird, J.R., Scott, M.D., Clayton, E., Russell, Geographic distribution and the characterization of L.H. and Cohen, D.D. (1980). PIXE–PIGME studies of volcanic glasses in . Archaeology and Physical artefacts. Nuclear Instruments and Methods 168: 447–452. Anthropology in Oceania 12(3): 173–201. Tuhinga 29: 90–164 Copyright © Museum of New Zealand Te Papa Tongarewa (2018)

In search of the North Island Archaic: archaeological excavations at Sarah’s Gully, Coromandel Peninsula, New Zealand, from 1956 to 1960 Janet Davidson

Honorary Research Associate, Museum of New Zealand Te Papa Tongarewa 9 Maple Lane, Martinborough, New Zealand 5711 ([email protected])

ABSTRACT: Excavations led by Jack Golson at Sarah’s Gully on the Coromandel Peninsula between 1957 and 1960 represented a landmark in New Zealand archaeology. A total area of about 1350 m2 was hand-excavated at six separate locations, recorded as three separate sites. Stratified deposits are dated to various points between aboutad 1300 and 1800. Artefactual remains show little change through time, with continuity in use of the important basalt resource at nearby Tahanga hill for stone adzes. Although moa and the North Island adzebill (Aptornis otidiformis) soon disappeared from the record, the faunal remains as a whole reflect continuity in fishing, shellfish-gathering and the taking of birds. The original forest was cleared very soon after people arrived. Storage pits, probably for kümara (sweet potato), were present throughout the sequence. Although the headland at the entrance to the bay was partly fortified during the middle of the sequence, the defences appear token in nature. Overall, the excavations revealed a way of life that involved periodic visits to fish, to garden, and to make and repair stone tools. This changed very little over the centuries. . KEYWORDS: Sarah’s Gully, archaeology, moa-hunting, material culture, faunal remains, chronology, storage pits, burials, regional sequence.

Introduction The present paper provides an account of the four seasons of excavation at Sarah’s Gully, for which there have previously been only interim and summary reports (Golson Cambridge-trained archaeologist Jack Golson arrived in New 1959a,b; Birks 1960; Birks & Birks 1970, 1973). My aim is Zealand early in 1954 to take up an appointment as lecturer to reconstruct the excavations from the notes, photographs in prehistory in the young Anthropology Department at and material items that have survived, and consider them in what was then Auckland University College. He immediately the light of more than five decades of subsequent research in immersed himself in New Zealand archaeology: its history, New Zealand archaeology and prehistory. literature, and field evidence (Golson 1957; Groube 1993). The paper draws on original field notes, including After relatively small field operations at Oruarangi and diaries and drawings made available by Golson and Lawrie Great Mercury Island (Ahuahu) (Golson 1955), and a rescue Birks; photographs in the archive of the Anthropology excavation at the volcanic cone of Taylor’s Hill in Auckland Department, , and photographs (Leahy 1991), Golson set about finding and excavating a site provided by Wallace Ambrose and Helen Birks; additional or sites in the North Island that would produce faunal and notes, plans and photographs assembled by Roger Green; cultural evidence for comparison with the large, extensively and the dating of surplus material from some radiocarbon excavated South Island moa-hunter sites, such as Wairau Bar, samples originally submitted by Golson for dating by the Shag River and Papatowai. The results of his excavations at Institute of Nuclear Sciences. This material is held in the Sarah’s Gully on the Coromandel Peninsula and Pig Bay on Auckland Institute and Museum. I took part briefly in the Motutapu Island near Auckland were fundamental to the final excavation season at Sarah’s Gully in January 1960. development of his views on what he would call the North In his quest for a suitable Moa-hunter (or, as he would later Island Archaic (Golson 1959a). call it, Archaic) site, Golson drew on artefact distributions Archaeological Excavations at Sarah’s Gully 91

and existing literature. His review of documented finds The scene was set for the subsequent four seasons of of stone adze blades in museum collections revealed a excavations at Sarah’s Gully and Opito. In his fieldwork concentration of stylistically early adzes on the east coast of on Great Mercury Island (Ahuahu), Golson had already the Coromandel Peninsula (Golson 1959a: 69). An early established the desirability of a location that would be report of moa bones in association with humans (Owen attractive to his volunteer labour force: a site adjacent 1848, cited by Duff 1956: 250), and a 1930s excavation to a safe swimming beach in a remote place with fishing that had yielded moa-hunter-style fishhooks and other opportunities from nearby rocks. Sarah’s Gully amply met items (Fisher 1936), drew his attention to Opito Bay on the these requirements. Although the location was accessible Kuaotunu Peninsula north of Mercury Bay. by way of the then notoriously dangerous Black Jack Road Golson went with a small group of colleagues to Opito from Whitianga and a walking track, the gear and some of early in 1956 and talked to one of the landowners, R.H. the personnel travelled from Auckland on the coastal vessel (Skipper) Chapman, who suggested he go to nearby Sarah’s Lady Jocelyn and were rowed ashore. Gully, where artefact collector Arthur Black had recently Sarah’s Gully and Opito are part of the rohe (territory) of made some interesting discoveries. During their visit, a Ngäti Hei. However, there is no indication in the surviving member of the group found a typically early-style ‘hog- documentation of the four seasons’ work there that any back’ adze that had eroded out of the beach face (Susan approach was made to the iwi, or that Skipper Chapman Davis, pers. comm., January 1959). had any involvement with Ngäti Hei.

Fig. 1 The location of Sarah’s Gully, showing sites recorded in the New Zealand Archaeological Association’s database (New Zealand Archaeological Association 2009): A, the Kuaotunu Peninsula on the east coast of the Coromandel Peninsula in northern New Zealand; B, Sarah’s Gully; C, Tahanga hill; D, Skipper’s Ridge and the Opito Beach Midden. The inaccuracy of the overlay reflects the difficulty of precisely locating sites using grid references. The overlay is slightly to the south of the true positions. 92 Tuhinga, Number 29 (2018)

The locality

Sarah’s Gully (Whaorei Bay) is a small northwest-facing bay on the northern part of the Kuaotunu Peninsula, between the larger beaches of Otama to the west and Opito to the east (Fig. 1). It has an open sandy beach, with rocky shores at either end. The main stream enters the sea about two-thirds of the way up the beach to the northeast; there is a smaller stream further up the beach and a third that comes down into the beach at the northern end, separating the northern headland from the remainder of the bay (Fig. 2). The original Sarah’s Gully site was initially recorded by Green as Sarah’s Gully Settlement, N40/9 (now T10/167), and is called the Settlement in this paper. South of the stream was the Cross Creek site (N40/260; T10/399), surface- collected by Ron Scarlett and others during Golson’s excavations, and later investigated by Brenda Sewell (1984, Fig. 2 Sarah’s Gully, looking northeast. The sites discussed 1986, 1988; Furey et al. 2008). in this paper are at the far end of the beach: 1, the Settlement; On the northern headland is the Sarah’s Gully Pä 2, Sarah’s Midden; 3, the Pä (photo: W.R. Ambrose). (N40/10; T10/168), here called the Pä. Below the Pä on the seaward side of the mouth of the northern stream was

Fig. 3 The excavations at Sarah’s Gully Settlement, as mapped by Roger Green in January 1959. Some minor inaccuracies in the map are discussed in the text. Details are shown as follows: 1, Fig. 12; 2, Fig. 13; 3, Fig. 15; 4, Fig. 16. Archaeological Excavations at Sarah’s Gully 93

Fig. 4 A view of the sites, showing the excavation of Area AB (photo: W.R. Ambrose). Sarah’s Midden (N40/13; T10/171), investigated during History of investigations the last season, and so named because its upper layer contained items of European material culture. According The Settlement occupied an extensive area of the foreshore to local legend, Sarah was an elderly Mäori woman who in the northern part of the bay and extended onto a ridge lived alone at Sarah’s Gully with a pet pig. She used to walk behind. Various areas were investigated over three seasons to Whitianga on Fridays with the pig trotting behind her. (Fig. 3). She eventually died at her home and the pig was accused of The first season ran from 28 December 1956 to 14 nibbling her body. January 1957 (Scarlett 1957: 3). It began with the excavation By 1990, the seaward edges of the Settlement had been of Area AB (Fig. 4), followed by Area C and an investigation affected by coastal erosion and one part (the coastal end of of the beachfront part of Area D. The second season, in Area D) had completely disappeared. The inland part of January 1958, extended the investigation of Area D inland, Area D was compromised by a small bach. The Cross Creek initially along the natural terrace beside the stream and then site had been seriously damaged by stock and wind erosion. up onto the ridge behind (Scarlett 1958) (Fig. 5). Sarah’s Midden could not be found and had, perhaps, been The third season, December 1958 and January 1959, completely eroded away (Sewell 1990: 197, 200, 202). completed the excavation of the upper part of Area D. Two days were spent investigating a midden (N40/3; now T10/161) at Opito Beach (reported by Boileau 1980). Green mapped both the Settlement and the Pä, and initial excavations were undertaken on the latter. Hamilton Parker began the excavation of Skipper’s Ridge (N40/7; now T10/165), a pit site immediately inland of the Opito Beach Midden.

Fig. 5 The extensive excavation in the inland part of Area D, towards the end of the second season (photographer unknown; image provided by W.R. Ambrose). 94 Tuhinga, Number 29 (2018)

During the summer of 1959/60, Golson was in New In the eastern part of the excavation – E and F, and part Caledonia. Parker continued excavations at Skipper’s Ridge of the G, lines of squares – the deposit was relatively shallow, (reported in Davidson 1974, 1975; Davidson & Green 1975). comprising only three main layers. To the seaward side, Birks directed excavations at the Pä and Sarah’s Midden. He however, there appeared to be several structures cut into and his wife had continued the excavation of the Pä in the the natural consolidated sand deposits and filled with a series early months of 1959, and took it further in early 1960 after of cultural layers. As this area had been severely cut back the conclusion of the main excavation season in January. by stream erosion, it was difficult to assess the nature of the Altogether, the excavations at the Settlement exposed a structures. little more than 1000 m2. The excavation on the Pä exposed The total depth of the upper deposits varied between an additional 332 m2. In contrast, only 6.5 m2 were fully about 28 cm and 38 cm (excluding overlying patches of excavated in Sarah’s Midden, with a further 3.25 m2 taken windblown sand). The layers were described as follows, to the surface of layer 3. Even so, this small excavation made from the top: an important contribution to the whole picture. 1 Black sand beneath sparse turf and windblown sand; quite hard in places, with some bone, stone flakes and The excavations broken oven stones. There was a restricted shell lens on The Settlement the seaward edge in E11. Remains of a small hearth or Area AB hängï were found in F13. This excavation was on a gently sloping natural terrace. 2 Light brown sand, with shelly patches in western parts The main stream had cut into this terrace, exposing cultural of the area. There were two distinct areas of hängï deposits. Amateur archaeologist and collector Black and an stones in G8 and G8/9 and one in Area B. Some associate had dug into the site from the seaward edge in cultural material. what became Area B, exposing deposits about 1.5 m deep, including what appeared to be a very large hängï (earth 3 Mainly darker brown sand, but varying in colour and oven) (Fig. 6). composition, sometimes softer, and sometimes grittier. A grid of 6 ft (c. 1.8 m) squares was laid out parallel to Layer 3 overlaid the natural consolidated sand the beach face and designated E–H from east to west (the throughout the eastern part of the excavation. Hollows underlay the layer 2 fire features; other shallow, stream edge), and 7–16 from south to north. Initially, an area amorphous features were found in the surface of the of 18 x 20 ft (c. 5.5 x 6 m) was excavated in Area A to the natural, filled with layer 3. There was a possible shallow south and 16 x 14 ft (c. 4.9 x 4.3 m) in Area B. Area B was posthole in E/F9 and two definite postholes in G9 and subsequently extended north by 6 ft (c. 1.8 m) and the baulk H9, 7½ in (19 cm) and 16 in (40.5 cm) deep, between the two areas was removed (Fig. 7). This resulted respectively. Remains of a hängï were found in E/F16; 2 in an excavated area of approximately 140 m , as the H row a similar dished depression in the surface of layer 3 of squares reached to, and sometimes just beyond, the edge in E/F15 contained black soil and charcoal but no of the eroding stream bank. stones. Some cultural material.

Fig. 6 The exposed beach section in Area B in December Fig. 7 The completed excavation of Area AB, looking towards 1956, after cleaning down and before excavation (photo: the Cross Creek site. A drain is clearly visible in the centre W.R. Ambrose). (photo: W.R. Ambrose). Archaeological Excavations at Sarah’s Gully 95

Fig. 8 The vestigial remains of a complex of structures on the seaward edge of Area AB. The upper deposits can be seen in section beneath a layer of white windblown sand at top right (photo: W.R. Ambrose).

In the baulk between Areas A and B was a drain, dug into these two squares is uncertain). In layer 2, higher up in the the natural consolidated sand and filled with layer 3 (Fig. baulk, were two substantial clusters of hängï stones in H11 7). This entered the excavation at the junction of E12 and and two smaller patches in H12. E13, ran diagonally through F12 and G12, and was lost in In the H line of squares in Area A were remnants of several the eroding beach face in H12, so that its relationship to structures cut into the natural consolidated sand. A ‘ledge’ presumed structures in this area could not be ascertained. In ran along below the edge of layer 3 in H9–11 and appeared January 1958, two squares (shown in Fig. 3) were opened on to have been cut by a deep pit, which in turn had cut into an the inland side of the Area AB excavation in an attempt to earlier pit. Complex fill layers suggested occupation breaks trace the drain upslope. The result was the same as had been between the two pits and between the pits and the shelf. found that summer in the inland part of Area D: the drain Several postholes were found, which appeared to relate to a petered out at the base of the slope. These squares added at possible house structure (Fig. 8). least 10 m2 to the excavation of Area AB (the exact size of Layers in the lower part of Area A were as follows (Fig. 9):

Fig. 9 The deep section in Area AB. The drawing is annotated ‘2 ft N. of line 10/11’ (photo: W.R. Ambrose). 96 Tuhinga, Number 29 (2018)

4 Brownish grey with shell fragments, extending into Area C H10 and thinly into H9. This was on the other side of the small stream north of Area 5 Similar to layer 2, with numerous shell fragments. AB. An 18 x 18 ft (5.5 x 5.5 m) square was laid out and excavated in four quadrants. The baulks were subsequently 6 Similar to layer 4. removed (Fig. 10). Green’s map (Fig. 3) appears to be in error in its depiction of Area C. There is no record of any 7 Grey sand with considerable cultural material. excavation other than the original square. 8 Reddish sand, apparently derived from the consolidated The same stratigraphy was found throughout this natural and apparently a living surface. excavation and was described in the northwest quadrant as follows: 9 Grey sand with considerable cultural material, more in the form of a midden than a living surface. 1 Topsoil.

10 Yellow shelly sand, above a natural deposit of white 2 Similar to layer 2 in Area AB; thick in the west, shelly sand. thinning to the east. 3 Dusty grey sand with charcoal. The first significant In Area B, layer 3 was thicker in G15, above what had been cultural material was a concentration of shells and presumed to be a large oven. This main feature, sectioned by artefacts at the base of this layer. Black, was found to be a large, deep, dished depression about 15 ft (4.6 m) wide, filled with very black soil and charcoal. 4 Damp grey sand with an almost solid spread of bones There were no stones in the depression, but there was a layer and flakes, suggesting a living surface. Various black of very large stones, almost like a pavement, on the centre of patches in and on the surface of this layer were its surface, with a much smaller, shallow, dished depression, sometimes designated 3A. Beneath this in places was filled with sand and charcoal, immediately beneath it, dug layer 4A, virtually sterile white sand with shell into the surface of the big depression (see Fig. 6). fragments. Most of the artefacts from the surface and layer 1 were 5 Compact black sand with broken oven stones and not catalogued to square. In layer 2, in addition to nine charcoal, which filled and levelled a large, shallow, ‘loose finds’, 10 were from the G row of squares, with one drain-like depression running across the quadrant. each from F8 and E14. Layer 3 saw a concentration in the area of G–H/9–12 with only scattered finds elsewhere. This Apart from the drain, the only features recorded in Area C continued to the bottom of the deposits, with the exception were a complete hängï, two patches of hängï stones, and of layer 4, with only a fishhook fragment and some stone a cluster of three or four very shallow postholes, all in the flakes from G–H/10–11. Finds in layers 7 and 9 were almost northwest quadrant, where the deposits were deepest and entirely restricted to G11. most productive of cultural material.

Fig. 10 The completed excavation of Area C, looking towards the beach. The principal section – the west face of the northwest quadrant – is visible at top right (photo: W.R. Ambrose). Archaeological Excavations at Sarah’s Gully 97

Fig. 11 The coastal part of Area D before excavation. The burial complex was in the vicinity of the three flax bushes to the right (photo: W.R. Ambrose).

Area D coastal A Windblown sand. Area D was near the point where the main stream entered 1 Black buried turf line. the bay. It was here that the hog-back adze had been found on the beach during Golson’s first visit (Fig. 11). During 2 Whitish sand with shell fragments, thick in the north, the excavations at Areas AB and C, various people had been thinning out to the south. collecting artefacts and faunal material from the eroding 3 Thin, damp, dark sand. and shifting surface here. A small exploratory trench on the crest between Areas A and D had proved shallow and largely 4 Thin cultural layer, sometimes black with charcoal, sterile. Black and Scarlett then started digging into a face a elsewhere sticky and clay-like. little further south, where there was a thick surface layer of A hasty trench was dug to link I and II, confirming that windblown sand, and encountered an artefact-rich deposit. layer 4 was the main cultural layer. A ‘hearth’ was found in Golson felt pressured to investigate further. Squares were laid F8, associated with dog, seal, moa and fish bones, as well as out in alignment with the grid of Area AB. Excavation began various stone tools. with squares E/F3 and E/G4 at the northern end of the When Black found a cultural layer with a second burial terrace (labelled as I in Fig. 12). Golson commented in his pit dug into it in the vicinity of F/G13 (Burials 7–9, labelled field notes that ‘finds were few’ in this area. The stratigraphy as C in Fig. 12), Golson extended along the lines of F–G/11– was as follows: 14 (labelled as III in Fig. 12), curving inland and downwards 1 Turf and blown sand. on to an area described as Platform E. Stratigraphy in III was the same as in II, although layer 4, rich in cultural material in 2 Black sand with two hängï. F11, petered out before the end of Platform E. Several more 3 Grey sand sloping to the bottom of the terrace. burials (1, 2 and 10, not marked on any existing plan) were found. 4 Puggy grey material, thought to be natural. Although The burials were very carefully excavated and cultural items were said to have been found eroding photographed in some detail, although most were not drawn. from this layer, none was found during controlled There appears to have been no other cultural activity in this excavation. area. At the request of the landowner, Skipper Chapman, Attention then turned to F–G/9–10 (labelled as II in Fig. 12), the human remains were reburied on the property. Golson where Black and Scarlett had already excavated, extending wrote in his diary, ‘Necropolis on Platform E does not really inland from their efforts. Parts of an eroding burial (Burial interest me’. A valuable opportunity to learn something of 3, labelled as A in Fig. 12) were found by Scarlett in G10. the life histories of these people, and perhaps of their genetic The burial pit had been dug through the ashy cultural layer. affinities, was lost. The burials are further described below Stratigraphy was as follows: and in Appendix 3. 98 Tuhinga, Number 29 (2018)

In the 1958/59 season, squares 1B, 2A and 2B (Fig. 12) Area D inland were dug to try to link up the stratigraphy of I and II. These The large area excavation running inland began on the flat revealed a sloping depression between the two areas, similar terrace immediately behind the seaward part of Area D (Figs to one found in square 5C (described below), which had 12 and 13). Stratigraphy on this terrace was as follows: contributed to the erosion. Both of these appeared to have 1 Turf. been formed by drains coming down from above (see Fig. 12). 2 Black sand, thicker towards the seaward end, containing It seems that there was only one main cultural occupation a few flakes of obsidian and pieces of burnt stone. layer in this area: layer 4, which was thin but relatively rich in both faunal and artefactual material. Transient later 3 Greyish sand, thicker towards the seaward end. This was thought to be the same layer that sealed the ‘moa- occupation is indicated by the hängï in layer 2 in I and a hunter’ deposit in the coastal excavation in Area D. few artefacts from upper layers. The fact that several of the Activity on the surface of layer 3 on the inland terrace burials had been dug into layer 4 from a higher level suggests consisted mainly of ‘fire pits’: shallow scoops with that the whole burial complex was probably more recent charcoal but little stone. than the layer 4 occupation. 4 Softish, more yellow sand, beginning in square 2 and The exact extent of the excavation in Area D coastal is not thickening towards the inland end, but giving out known. Green’s map, made two summers later, following towards the creek and the rise up the ridge. further erosion, is clearly not quite correct, and there is no 5 More compact, greyish sand with minute shell. documentation of the linking trench between I and II or of Restricted to the central part of the terrace, from 2 the burial area. A conservative estimate of 35 m excludes square 4A to square 8. Scattered mammalian bone and the burial area. some charcoal pieces were present in square 7A. Natural Hard consolidated sand.

Fig. 12. Plan 2, Area D, showing the further extension of excavations inland along the low terrace; the upper termination of the second and third drains; and the initial extension up on to the flat above. Archaeological Excavations at Sarah’s Gully 99

Fig. 13. Plan 1, Area D, showing the area of the coastal excavation and the initial extension inland, including parts of the first two drains. A. Burial 3; B. Burial 4; C. Burials 7 to 9. D. Burial 11. The other burials were in the unmapped area just to the south.

The principal features here were three shallow but distinct square 5B petered out before what appeared to be a natural drains, first encountered in square 5 (Figs 12 and 13). They embayment leading down to the stream on the western were dug into the underlying natural, sealed, and sometimes edge of square 5C. filled with layer 5; in squares 4A and 5, the northern edge Other structural features were shallow postholes in of the first drain marked the northern edge of this layer. several squares, apparently dug from the surface of layer In squares 8A and 9A, where layer 5 was not present, the 4, notably a line of four in squares 4 and 4A. A small drain was filled with layer 4. The drains petered out at the rectangular feature in square 4A was deemed to be a modern base of the slope on the uphill side. The one traced into intrusion. 100 Tuhinga, Number 29 (2018)

Fig. 14. Burial 11 in square 5A Area D. Upper: photograph W.R. Ambrose; lower: drawing by John Parry.

A remarkable burial (here designated Burial 11) was found hängï in the natural surface. Half squares 9d and 9e (not in square 5A (Fig. 14). There was no grave pit and the body shown on Fig. 13) were opened to carry the excavation to appeared to have been placed on the natural ground surface the flat area above. Here, layer 2 became more turf-like and and covered with heaped layer 5 material. layer 3 more clay-like but both still contained sand. The excavation was then extended to the base of the On the flatter area above, in squares f–g/9–14, the slope through squares 8 and 9a–c, where the stratigraphy stratigraphy was described as follows: was limited to the turf layer (equated to the previous layer A Topsoil. 2) and a greyish to lighter-coloured sand (layer 2, equated to previous layers 3–4) on top of consolidated natural sand. B Soft greyish-yellow soil of sandy texture. There were traces of hängï and burning in layer 2 and actual C Friable brownish-yellow soil. Archaeological Excavations at Sarah’s Gully 101

The principal features in this area were pits (Fig. 15), which to confirm the link between the lower flat and the pit were covered and sealed by layer C. The pits are described complex by the excavation of squares 12b–e. This did not below. A large, deep posthole in square 9g was thought prove satisfactory, however, as layer C was not found. The possibly to be for supporting a pätaka (raised storehouse) excavation then progressed to 11–12/h–i and intervening or whata (storage platform). Squares 7f, 8g, 10i (not shown baulks, where two further shallow rectangular pits (6 and on the plan), 12j–k, 13i–k, 14i–k, 15k and 17f (not shown 7) were found (Fig. 15). Squares above and inland of these on the plan) were sterile. There are no notes about other pits were sterile, except for 11k, until the row of squares featureless squares in Fig. 15, which may also have been sterile. 12–16m was reached (Fig. 16). In square 11k, a complex What Golson considered the crucial stratigraphic link structure of what had probably been superimposed hängï between the pits and the lower terrace was found in square pits was discovered. Only some contained stones, although 13c, where layer C, which sealed the pits above, was found most or all contained charcoal. This feature was thought to underlie layer 3, followed up from the flat below. Golson to be more recent than most of the features further up. In (1959a: 45) argued that this proved the pits were associated squares 12–16/m–p, three layers were identified: 1 and 2 as with the ‘moa-hunter’ occupation on the beach front before; and an underlying cultural deposit designated layer below, which was also sealed by layer 3. X, which contained cracked hängï stones, charcoal, a shell In the summer of 1958/59, a further attempt was made fragment and a few pieces of obsidian.

Fig. 15. Plan 3 Area D, showing the pits uncovered. The feature in square k11 appeared to be a complex of superimposed hängi pits. The large posthole in the corner of square G9 may have supported a raised storage house (pätaka) or platform (whata). 102 Tuhinga, Number 29 (2018)

Fig. 16. Plan 4 Area D, showing the features (mostly postholes) in the highest area investigated. Numbers refer to features described in the text.

The principal features in this uppermost area were although this was not carried out. postholes, some definite, some doubtful (Fig. 16). Although The exact area of the extensive excavation in Area D it is possible to see a rectangular grid pattern in these, their inland is not certain, as at least two squares are not shown on variation in width and depth leaves this interpretation the plans and several apparently excavated during the second uncertain. Of the larger features, 13 was described as a season, before the map was made, have no documentation. ‘shallow oval pit’, probably filled with layer B; 35 as a ‘large Assuming that these were excavated, the total size of Area D hole’; 39 as a ‘pit’; and 46 as a ‘not very convincing hollow’. inland was 786 m2. Most significant was feature 37, a definite pit sealed by layer X and partly overlain by a hängï associated with layer B. Pits and drains Thus, there was some evidence for successive use of this Details of pit dimensions and features are given in Appendix upper part of the site, as there had been below, with sporadic 1. Pits A, B and C were very shallow and ill-defined. Pits 3–5 cooking activities overlying earlier features. Although were larger, more clearly defined and slightly deeper. Pits 1 numerous charcoal samples were taken from this area, none and 2, which might be described as bin pits, were smaller and was dated. Similarly, numerous soil samples were taken for deeper, with a ledge below the surface on some or all sides. analysis of the composition and origin of the various layers, Pits 6 and 7 were again very shallow. All, including the bin Archaeological Excavations at Sarah’s Gully 103

pits, had lower edges on the downslope sides and/or ends. mandible, while the other two were represented only by the Some of the postholes in the floors of the rectangular pits skull and some vertebrae and ribs. Scarlett noted that the were quite deep, indicating a fairly substantial superstructure. ‘base of the skull’ had been removed in some instances. He There were no signs of fire features or habitation debris in found no evidence of trauma, but the incomplete nature of any of the pits, most of which were too small to have been many of the burials does not rule this out. An example of anything other than storage structures. Several had a distinct severe head trauma and decapitation was reported from a site thin layer of sand or other introduced material on their floors. at nearby Opito (Calder & Calder 1977). Houghton (1977: The two rounded features dug into the base of the 46–47) described this last individual as a female aged 20–25 downslope wall of Pit 3 were initially thought to be sumps years and her death as possibly ‘an anthropophagus occasion’ but, as they were not fed by the drain, seem more likely to (cannibalism). have been for the storage of some special part of whatever Three of the burial pits had been dug through the early was stored in the pit. occupation layer (4) in Area D into the sterile sand below. The drains across the pit floors, although shallow, were The others were beyond the limit of the occupation layer, clearly defined and were sometimes deep where they passed but since they were all in the same general area, it is likely out through the pit walls. However, the drop along the that they were contemporary – the remains of one group of drain from Pit 5, from the point where it exited the pit to people, who used this place for burial over a relatively short where it passed into the baulk 10g/h, 12 ft (3.7 m) away, period time. This time could have been at any point after was only 4 degrees. the deposition of layer 4 in about the fourteenth century. The drains on the terrace below were wider than the The absence of any European items does not rule out the pit drains but just as shallow. Attempts to follow drains possibility that they belong to the early European contact either upslope from the lower terrace or beyond the limits period, as there were no grave goods of any kind. The of excavation on the middle terrace (e.g. in part square 9h) incomplete nature of most of these burials raises questions were unsuccessful, and Golson regretfully concluded in his that cannot be answered. Were some exposed bodies subject field notes that the drains gave out at the tops of the steeper to scavenging? Were some bones deliberately removed before slopes and resumed at their bases. As noted above, this also burial? Why were some cranial bases removed? Or was the appeared to be the case with the drain in Area AB. incompleteness at least partly due to erosion and fossicking? This last could, at best, be only a partial reason, given the apparently undisturbed nature of Burials 7–9, for instance. Burials Scarlett noted evidence of tooth wear (including so-called The burials from the southern part of Area D (Fig. 12), ‘fern-root’ wear in one case). The most detailed description uncovered during the first season of excavation, were is of Burial 2, who had severe caries, pre-mortem tooth described by Scarlett before reburial. An edited version of loss and considerable wear on the surviving cusps. All this his report is presented as Appendix 3, and the comments suggests a harsh diet. attributed to Scarlett here are from that report. Scarlett noted The burial found further inland on the lowest terrace that the crania were typical of the Polynesian form, although of Area D (here designated Burial 11) was quite different. he did not have the necessary equipment to take precise and This person was lying on their right side. Both lower legs detailed measurements. Some of the burials were eroding, were bent upwards towards the pelvis, and the hand and foot some had already been partially uncovered by Black, and bones were scattered. Both arms were folded, so that the some were not fully excavated; even so, this does not explain hands could have been on the breast. There were no grave the incompleteness of most of the skeletons. goods, unless a ‘fish bone’ (with no modification) shown at All were crouched and/or bundle burials, including one the neck in the drawing was a snapper tilly bone worn as an group of three that Scarlett considered possibly members ornament (one such perforated ornament was found in layer of one family (parents and one child). They varied from 2 of Sarah’s Midden). virtually intact trussed bodies to bundles of disarticulated An extended burial was also found at the Opito Beach bones, suggesting that some had been buried immediately Midden, during the brief excavation there (Fig. 17). It was after death, while others had been exposed before retrieval fairly well down in the sandy matrix, with no indications of and burial. Skulls and/or mandibles were missing from some a grave pit. The body had been placed on its left side, with otherwise largely complete individuals, while in the multiple the legs slightly flexed. The arms appear to have been folded. burial of Individuals 7–9, one was complete except for the There were no grave goods. 104 Tuhinga, Number 29 (2018)

A cache of 14 large, unfinished adzes found with a burial at Opito many years ago and now in the Auckland War Memorial Museum (Auckland Museum), and a burial with two necklaces at Hahei (Edson & Brown 1977), are evidence that in this area at least some early people were buried with precious things. However, the absence of grave goods at Sarah’s Gully need not mean that none of the burials is early. Although grave goods are more typical of earlier practice, by no means all early burials had them. Sarah’s Midden Sarah’s Midden was on a natural terrace on the western side of the northernmost creek, below the headland on which the Pä is situated. Two 6 ft (1.8 m) squares, designated 100 and 101, were excavated over several days in January 1960 (Figs 18 and 19). A third square (102) was fully excavated only to the base of layer 2, with 1 ft (30 cm) square test pits taken down to natural in the four corners. The excavation Fig. 17. Burial 12, from the Opito Beach Midden. has been mentioned only minimally in print (Birks 1960: 20; Houghton (1977: 48–51) described two skulls from Green 1963a: 67–68). Sarah’s Gully (Individuals 4 and 5), presented to the Anatomy Department at Auckland University by Golson, which do not match any of the burials described by Scarlett. Since all the human remains from the coastal part of Area D were reburied, it is possible that these two skulls are from Burials 11 and 12. Houghton described Individual 4 as a female aged about 25. She had little tooth wear, moderate subgingival calculus and no caries. This is in marked contrast to Scarlett’s comments on the teeth of the individuals in the main burial area in the coastal part of Area D. Burial practices were variable at all periods of New Zealand prehistory, although there was a trend away from primary burials in and near settlements during the early period, to seclusion of human remains removed from living areas in later times. Extended, crouched and disarticulated burials are found in both early and later sites (Davidson 1984: 172–177). All that can be said about the Sarah’s Gully burials is that Burials 1–10 appear to reflect repeated use of one small area by one group of people, probably over a fairly short period, which could have been any time after the early layer 4 occupation of the coastal part of Area D. Burials 11 and 12 could be of any age, and may be representative of a different group of people or of a different period of time. Houghton (1977: 52) used a now discredited technique of nitrogen dating to argue that Individuals 4 and 5 were relatively early in the regional sequence, and then relied on this interpretation to support his theory that early people in the area had a softer diet than those of the later pre-European period. Fig. 18. Birks’ plan of the excavation of Sarah’s Midden. Archaeological Excavations at Sarah’s Gully 105

Birks’ notes on this excavation include the site map (Fig. Stratigraphy was as follows (Fig. 20): 18) and a sketch of features in layer 4 of square 101, the most 1 and 2 Turf and a soil layer. Thin lenses of sand separated productive of the three. He described stratigraphy, features layer 2 from layer 3 in several places. and finds for each square. There are section drawings of all four sides of square 101 and two for each of the other 3 Mixed blacker sand and soil. squares, as well as a drawing of one face of square 101 in 4 Yellowish-brown sand. a different hand. In addition to Birks’ photographs, I have consulted several that were taken by Anne Leahy and me. 5 Darker sandy material, but not as consistently black as layer 3. Natural Clean sand in square 102; clean beach sand interspersed with water-worn stones and boulders in square 101.

Layers 4 and 5 petered out at the base of the slope in square 100. Photographs suggest that layer 3 was a dark sandy layer with relatively little cultural material compared with layer 5. They show a large shallow hängï at the base of what is presumably layer 3, from which radiocarbon sample WK35393 has been dated. Birks described this as the largest and uppermost of three successive hängï pits in this part of layer 4 in square 101. There is also a dated sample (WK35394) from a hängï supposedly associated with layer 5. Moa bone was definitely associated with cooking in layer Fig.19. The excavation at Sarah’s Midden, January 1960. 5 (Fig. 21). Two postholes were recorded at the base of Trevor Hosking and Bob Jolly are excavating in square layer 3 in square 101, which penetrated through layer 4 101; square 100 is being turfed. Photograph J.M. Davidson into layer 5.

Fig. 20. The east face of square 101, Sarah’s Midden. Photograph L. Birks. 106 Tuhinga, Number 29 (2018)

The Pä

As noted above, excavations on the Pä took place over two seasons. In January 1959, an initial line of squares was laid out through the defensive ditches and down the central part of the site, with a row at right angles sectioning the southeastern scarp. The grid was 12 ft (3.66 m), within which 9 ft (2.74 m) squares were excavated (Golson 1959b: 16–17). Initially, squares 10, 12, 14–18 and 18D–E were opened. The Birks returned in mid-February and carried on by themselves until mid-April (Fig. 22). During the main season in January 1960, they resumed work, and later continued by themselves until they had completed all the areas shown in Fig. 23. According to surviving plans and notes, the total area excavated was 327 m2, excluding some test pits outside the defences. Documentation includes notes by both Golson and Birks on almost all squares and baulks excavated; notes by Birks on most but not all sections and plans; and numerous black- and-white photographs, most of which have the context noted beside the mounted prints. The excavations were well summarised by Birks (1960). Although Green’s map (Fig. 23) shows the main axis of the Pä running from north-northeast to south-southwest, Birks in his notes describes the outer ditch as east and the of the Pä as west. Green’s orientation is followed here, with the defensive ditches running from west (Birks’ north) to east Fig. 21. Trevor Hosking uncovering part of a moa bone (Birks’ south). in Layer 5 of Sarah’s Midden. Photograph J.M. Davidson.

Some European material (mostly fragments of iron nails) was found in layer 2 in square 102 and layer 3 in squares 101 and 102. Birks considered that layer 3 was associated with Sarah’s occupation, and suggested that some square-shaped postholes probably represented foundation piles of her house. The significance of Sarah’s Midden, not recognised at the time, lies mainly in the faunal material (mostly from square 101), discussed below. The artefactual material was relatively sparse, although the presence of two dorsoventrally perforated bone trolling lure shanks is significant.

Fig. 22. A view of the Pä towards the end of the first season of excavation. Photograph L. Birks. Archaeological Excavations at Sarah’s Gully 107

Fig. 23. The excavation of the Sarah’s Gully Pä. The total area excavated has been superimposed on Green’s original map, which showed only the initial squares excavated in January 1959

Stratigraphy hängï; flakes and chips of basalt, chert and obsidian; shell; fish bone; and charcoal. There were some quite large pieces of Throughout the interior of the Pä, there were three main rock, sometimes wedged in features (Fig. A2.2) and sometimes layers: set in the ground (Fig. 24). There were also concentrations 1 Turf zone (generally less than 5 cm thick). of small smooth pebbles, often described at that time as moa crop stones or gizzard stones. Birks suggested in his notes 2 Developed soil zone (varying from 5 cm to 15 cm that a particularly dense concentration associated with hängï thick, but up to 30 cm thick on some slopes). stones in layer 3 in square 14A might reflect the cooking of a whole moa on the site. 3 Variously described as the weathered or puddled surface of the natural, not always present, and usually less than 8 cm thick except over features. Birks’ notes summarised the variable subsoil as ‘sand with rhyolite lumps, having the appearance of sandstone, overlying rhyolitic sand, having the appearance of intractable clay. Weathering and human interference had removed the upper component in many areas.’ In a few places, compacted fine sand, apparently derived from natural, sealed some features and was described as layer 4. Other localised layers were identified in some pit fills and in the ditch and bank defences. Cultural material, widely scattered through layer 2 in Fig. 24. Upright slabs in square 19A of the Pä. Photograph particular, included heat-fractured stone from the numerous L. Birks. 108 Tuhinga, Number 29 (2018)

The defences As shown on Fig. 23, the visible defences consisted of two discontinuous ditches, which defined the inland edges of the site, and an extensive but discontinuous scarp along much of the eastern (and less steep) flank, with a ditch and low outer bank along its southern part. The bank inside the outer ditch continued across to the northeast end of that ditch, although this was not shown on Green’s map (Fig. 23). The outer ditch was still present in square 10a. It appeared as if it had taken advantage of a natural dip at this point, as the inner side was slightly higher. A series of test pits (not marked on the plan) found that this ditch and a minimal inner bank continued across the entire width of the headland. The slight ditch appeared to continue down the western slope. In square 10, the outer ditch was only 20 cm deep in the west face and 46 cm deep in the east (Fig. 25). In the east face of square 10A, its base was 50 cm below the surface and 81 cm below the probable ground surface at the time it was Fig. 25. The outer ditch of the Pä in square 10, looking east.. Photograph L. Birks. dug, while the top of the bank was a mere 23 cm above that surface.

Fig. 26. The east faces of the inner ditch of the Pä in square 12 and the outer ditch in square 10.

In square 12a, the base of the inner ditch was 61 cm The two rectangles in squares 10C and 11C showed only below its present surface and had probably been dug to 84 what appeared to be natural slopes. However, the northern cm below the ground surface at the time (Fig. 26). The end of the ditch along the eastern side of the Pä was found present top of the bank was 46 cm above the original ground in the southeastern corner of square 13C, well beyond the surface; the bank had apparently eroded both back into the point at which it was shown on Green’s map. The other end ditch and for a considerable distance inwards. The inner of this eastern ditch was defined in the half square 25B. ditch clearly terminated at the edge of square 12. The only features suggesting any kind of structure along The layers in the banks and ditch fills included layers 2 the defences were a single posthole 30 cm deep in the crest and 3, and material derived from, and eroding back into, the of the bank associated with the outer ditch in square 10A; ditches. two very shallow square postholes on the bank in square 10 Archaeological Excavations at Sarah’s Gully 109

(Fig. 25); two matching, similar square postholes in baulk the site, a hard, hummocky ridge. Features such as storage 10–10a; and two 18 cm-deep postholes in the outer slope and hängï pits were clustered on either side of it, particularly of the bank at the end of the inner ditch in baulks 12–12a in squares 14–16A to 14a–16a (Fig. 28). and 13–13a. Birks (1960: 19) suggested that the square Most of the pits are described in Appendix 2. The majority holes on the outer bank perhaps related to an entrance, were relatively shallow rectangular pits, usually with central and represented a structure or structures set on, rather than postholes and sometimes also side postholes. They were not in, the ground. There were two rows of small, shallow unlike the pits in the inland part of Area D in the Settlement, postholes underneath the bank of the inner defence, perhaps although none had drains. There were also bell-shaped pits, representing earlier light fences. usually circular in plan with undercut sides. In the baulks The outer ditch and bank did not constitute a strong between squares 15 and 16, and extending into both squares, defensive feature. Although the inner ditch and bank were were various shallow intercut features, which may have been more impressive, they extended less than halfway across the parts of pits or, as Birks described them in his notes, ‘shelves’ width of the Pä. Moreover, the inner ditch and bank were into which oval or bell-shaped pits had been dug. rounded off in such a way as to suggest that there was no The numerous hängï were mostly associated with layer intention of continuing them further. It is hard to escape the 2, although a few were associated with layer 3. No clear conclusion that the defences of the Pä were largely symbolic. posthole patterns could be discerned, with three exceptions, where a circular scoop (presumed to be a hängï pit) was surrounded by a ring of postholes: nine ranging from 8 Other features cm to 19 cm deep around a small scoop in square 16a; six Other structural features included several varieties of pits, ranging from 15 cm to 29 cm deep around a larger scoop in assumed to be for storage, hängï scoops and postholes. square 15a, with traces of three others, largely eroded away, Many of the postholes contained stones wedged into them, in the edge of one side; and six ranging from 6 cm to 10 presumably to hold posts in place. cm deep around a small scoop in baulk 14–15. Birks (1960: Only two features were found in the area between the 18–19) speculated that these may represent some form of inner and outer defensive ditches (Fig. 27): a ‘grave pit’ in the wind protection near the exposed top of the ridge, noting northwest part of square 11 (Fig. A2.6), aligned parallel with the absence of similar posthole rings around hängï on the the outer ditch and bank; and a rectangular pit lying largely more sheltered landward side of the site. in baulk 11–12A. A single posthole was noted in square 11. The main group of features, both pits and hängï, began There was much more evidence of occupation inside the under the inner defensive bank in square 12a and continued inner ditch. The initial line of squares ran down the spine of to 16a. To the east, the concentration was mainly in squares

Fig. 27. The area between the ditches of the Pä, with relatively few features. Photograph L. Birks. 110 Tuhinga, Number 29 (2018)

Fig. 28. Features in the central area of the Pä. The pits are described in Appendix 2.

14–16, continuing in the eastern baulks between those level structures, other than the posthole rings noted above. squares and the A row, and into square 15A (Fig. 28). Although the numerous hängï scoops and extensive amounts Three aligned pits on the eastern slope appeared to be quite of heat-fractured stones gave an impression of considerable separate (Fig. A2.4). The large hängï scoops were almost occupation, this appears to have been largely transient in all in the 14–16 to 14a–16a area, although smaller scoops nature – at most, a few seasons of pit storage, and probably and clusters of hängï stones were scattered throughout more frequent periods of camping, cooking and perhaps the excavation inside the inner defence. There were few fish-drying. The many fires may even have been intended or no features in rows 17–19, except for the pit in square to signal a site more fully occupied than it actually was. 18B. Cultural material, including obsidian flakes as well as spalls from cooking stones, was scattered over a larger area, The sequence beyond the central cluster of features. Birks (1960: 18) summarised the sequence at the Pä from A small oval hollow in the southwest corner of square 18a early to late as follows: contained some very fragile human bones and three flakes Phase I. Marked by the occurrence of two pits, bell- of obsidian. Birks noted that this was more than 18 ft (5.5 shaped and comparatively deep, 14 feet apart near the m) from any other feature. There was very little evidence highest point of the site. It may include also four others, of occupation in squares 18a–18B. The bones included close by, much smaller, but generally similar in design. vertebrae, pelvic bones, one arm bone, several foot bones Two further pits, quite different in appearance from those and a tooth. There was pronounced marginal osteophytic of the other side, may nevertheless be contemporaneous, lipping of the vertebral bodies in the lower back area. since they have no obvious connection with either of Nothing was found to indicate any definite ground- the later phases, and in common with one of the larger Archaeological Excavations at Sarah’s Gully 111

pits, one had a layer of brown soil on the floor. Narrow rectangular pits (Pit A and Pit X) were covered by the bank in plan and straight-sided, they had a grave-like of the inner defence (Fig. 29) appearance, but the fill, clearly artificial, contained Material culture from the Pä was almost entirely of stone, nothing except scattered charcoal and the ubiquitous consisting of adzes and adze fragments, files and grindstones, flakes of andesite oven-stone. and a variety of flakes. Two bone items are described below; Phase II. To this belongs a number of rectangular pits, two others are simply fragments of worked bone, one from with two exceptions, shallow in relation to floor area, layer 2 and one from layer 3. and having a fairly regular pattern of associated postholds [sic]. Generally there were three or more on the longitudinal axis, and varying numbers up to four Chronology along one or both sides. The regular shape of some of Radiocarbon dates from the three excavation sites are these pits has been modified by later extensions. listed in Table 1 and depicted in Fig. 30. Golson obtained Phase III. Characterised by haangi pits over a large part five radiocarbon dates for the Settlement on unidentified of the site, notably on the more level central portion, charcoal. In 2008, the residues of the dated samples, held by and by a liberal dispersal of material from these, the Institute of Nuclear Sciences at Gracefield, Lower Hutt, including stones of local andesite, charcoal, shell, fish were examined by Rod Wallace, who was able to select pieces bones and flakes of obsidian. The pits in several cases of short-lived species from NZ356 and NZ357. These were had been dug into the fill of the earlier rectangular dated by the Radiocarbon Laboratory (WK20306 structures, but some of the largest had penetrated and WK20307, respectively). There was no residual sample through the topsoil into the underlying rhyolitic of NZ359, while the residues of NZ355 and NZ358 were material. deemed unsuitable for dating, the latter because it was part This sequence was most clearly expressed in square 14a of a post that appeared to consist of kauri (), and adjacent baulks 14a–15a. Here, a large bell-shaped pit which might have considerable inbuilt age (L. Furey, pers. (Pit A1) was overlain by a rectangular pit (Pit A) and three comm., 2009). The calibrated result suggests that this was hängï scoops, and four postholes ranging in depth from 6 in almost certainly the case. The close similarity between the (15 cm) to 26 in (66 cm) (not shown on Fig. 27) had been original and Waikato dates (Table 1) suggests that the original dug into the fill of Pit A. In squares 12a–13a, two adjacent results for NZ355 and NZ359 may also be reliable.

Fig. 29. The earlier bell-shaped pit A1 overlain by the rectangular pit A in squares 12a to 13a. 112 Tuhinga, Number 29 (2018)

Fig. 30. Calibrated radiocarbon dates from the three Sarah’s Gully sites.

The samples dated as NZ356 and NZ357 were both Table 1 Radiocarbon dates from the three Sarah’s Gully sites described in Golson’s submission as from layer 3, the former (CRA = conventional radiocarbon age). associated with fish bone and the latter associated with moa bone. However, it is clear from Golson’s diary that NZ356 Lab no. Context CRA was from a fire feature cut into the surface of layer 3, while Sarah’s Gully Settlement, T10/167 NZ357 was from the large oven in the surface of layer 4. NZ355 Area D, square F7, layer 4 656 ± 40 The two samples from Sarah’s Midden were labelled as being from a hängï in layer 3 and a hängï in layer 5, dug NZ356 Area A, square G9, hängï in surface of 216 ± 39 into the underlying natural sand. This supposedly layer 5 WK23206 layer 3 189 ± 30 date is anomalous, in view of the fauna in layer 5. Trevor NZ357 Area B, squares G15 and G16, hängï in 702 ± 40 Worthy (pers. comm., n.d.) considers the layer 5 deposit in WK23207 surface of layer 4 703 ± 34 Sarah’s Midden to be probably the earliest excavated deposit NZ358 Area A, posthole D, layer 8 873 ± 40 at Sarah’s Gully. The dated sample may therefore have NZ359 Area A, square H11, layer 9 650 ± 50 been wrongly labelled and actually be from one of several successive hängï dug into layer 4. Sarah’s Midden, T10/171 Birks obtained five dates for the Pä. As part of the project WK35393 Layer 3 hängï 285 ± 31 of redating ‘early’ dates from Sarah’s Gully, the residue of WK35394 Layer 5 hängï 369 ± 37 NZ1080 was examined by Wallace but was found to have Sarah’s Gully Pä, T10/168 no suitable material for further dating. The charcoal dates NZ1080 have been recalculated since the original determinations Charcoal from pit, phase 1 703 ± 46 were published, while the shell dates are the conventional NZ1081 Charcoal from pit, phase 2 388 ± 49 radiocarbon ages (CRAs) before adjustment for the marine NZ1082 Charcoal from defensive bank material 335 ± 48 effect. In Fig. 30, these two dates have been calibrated using NZ698 Shell from pit fill, phase 3 260 ± 51 Marine 13 with Delta-R set at 0 (C. Prior, pers. comm., 6 NZ699 Shell from pit fill, phase 3 292 ± 41 July 2015). Archaeological Excavations at Sarah’s Gully 113

The dates place occupation of layer 4 of Area D and layers currently accepted date of human settlement, and other early 4–9 of Area AB of the Settlement broadly between ad 1280 dates from the region, in-built age might be suspected. and 1400, as would be expected from the associated faunal Subsequent investigations of sites at Opito include and material items. These deposits can be grouped together Bellwood’s (1969) excavation further up Skipper’s Ridge, as representing an early phase of occupation at Sarah’s Gully. which gave two late dates, and the recent excavations on The age of the earliest deposit at Sarah’s Midden is unclear. a very large ridge site at the north end of the bay, which Unfortunately, the early date from the Pä could not revealed evidence of food storage, cooking and at least one be verified on identified short-lived plant species. It falls substantial house at various times in the late fourteenth, in the same time bracket as the early occupations at the sixteenth, seventeenth to eighteenth, and eighteenth Settlement, and it is quite possible that people in the area centuries (Bickler et al. 2014: 169). As was the case at Area at that time would have constructed a storage pit on the D inland at Sarah’s Gully, this ridge site produced relatively headland, particularly given the likelihood that at least some little in the way of faunal or artefactual material. of the pits in the inland part of Area D of the Settlement Taken together, the radiocarbon dates from Sarah’s were thought to have been contemporary with the early Gully and Opito cover the entire sequence of pre-European beachfront occupations. occupation in New Zealand. The remaining dates can tentatively be divided into middle (Pä Phase II) and late (Pä Phase III, Area AB, layer 3). As the only indication of European influence (apart from Material culture the material in upper layers at Sarah’s Midden) is a single The excavations at Sarah’s Gully did not produce the range of clay pipe stem from Area AB, it seems unlikely that any of material culture Golson had hoped for to assist his definition the deposits at the Settlement or the Pä is later than about ad of the North Island Archaic. The principal items recovered 1800. from the Settlement were fishhooks and items associated Sewell (1988: 8) obtained three radiocarbon dates for with their manufacture, stone adzes (largely broken and/or the Cross Creek site, just across the creek from Area D. partly reworked preforms) and a range of other stone items The sequence there has more recently been modelled using – mostly flakes, of obsidian, chert and basalt. Even so, most of the original dates for cultural layers 3A, 5 and 7, and four these objects are not incompatible with material from ‘early’ additional dates for layer 9 (Furey et al. 2008). The Kaharoa sites. No artefacts were found with the burials. The material ash, an important geochronological marker that has been culture from the lowest layer of Sarah’s Midden is even more dated to ad 1314 ±12 (Hogg et al. 2003), separates layers limited. 7 and 9. Unfortunately, the Kaharoa ash was not identified A few items from the upper layers of Sarah’s Midden, at any of the other Sarah’s Gully sites. However, both the Pä, and Area AB of the Settlement suggest later pre- layers 7 and 9 at Cross Creek fall within the time range of European visits to Whaorei Bay. the early-phase deposits from the Settlement and, at least potentially, at the Pä. The date range for the next cultural layer at Cross Creek, layer 5, has a slight overlap with the Personal ornaments early-phase dates at the Settlement, but contained shell fishhooks, none of which were found at the Settlement, and The small number of ornaments and worked bone (other probably corresponds more closely with the middle-phase than that relating to fishhook manufacture) at the Settlement occupations at the other sites. was disappointing. Of particular interest was ‘the poor The two sites at nearby Opito excavated as part of man’s rei puta’ (Fig. 31E), a pendant of Päua shell (Haliotis Golson’s research programme also date at least partly to the sp.), which has previously been described in some detail early phase. The Opito Beach Midden has a CRA of 689 (Davidson 1986). This came from the lowest cultural layer ±40 cal. bp (calibrated age at 2σ ad 1287–1395), while the in Area AB: layer 9 in square H11. From the same context date obtained for the earliest occupation at Skipper’s Ridge are a tiny bone needle, broken at the perforation (Fig. 31A); was reported as NZ1740: 780 +60/50 (Davidson 1974). 24 units of Dentalium nanum; and a single bird-bone bead. This result has been recalculated as 814 ±70 bp (University The apparent association of the pendant with Dentalium of Waikato n.d). This sample, from the base of a large units suggests that it might have been the central part of a storage pit, was identified as short-lived charcoal (Panax sp.), necklace. The only other definite item of adornment found possibly part of a post or roof beam. As it falls outside the in Area AB is a broken worked tooth of a dog or seal (Fig. 114 Tuhinga, Number 29 (2018)

31B), from layer 1. Probable parts of ornaments are a carefully cut, but unperforated, piece of a small sea mammal tooth (24 x 12 mm) from the surface, and a rectangular shell tab (29 x 12 mm) from layer 6 (Fig. 31F). The latter is clearly broken at one end and would have been longer. The only personal ornament from Area C (from layer 5) is a broken bone pendant, apparently made from part of a large, one-piece bone fishhook (Fig. 31D). The beachfront section of Area D produced two bird- bone beads and a cut section of bird bone, possibly relating to their manufacture, from layer 4 (Fig. 31C). Dentalium shell units are mentioned by Golson (1959a: 14), but are no longer present in the assemblage. A broken snapper tilly bone, perforated for suspension, from layer 2 of square 101 at Sarah’s Midden, is the only Fig. 31. Ornaments from Sarah’s Gully Settlement. A ornament from that site. No items of adornment were found bone needle; B worked dog or seal tooth; C bone tubes at the Pä. and worked bone; D broken bone pendant; E shell pendant; F Shell tab. Fishing gear

Parts of bone fishhooks and items relating to their Examples of one-piece hooks from each area, along with manufacture were found in Areas AB, C and the coastal part tabs relating to their manufacture, are shown in Figs 32–34. of Area D of the Settlement, along with a single example All are simple hooks, most with incurved points. Some from layer 3 in the baulk between squares 2 and 3, just inland have bait notches on the point leg, and one a notch on the from the latter. Nearly all are parts or preforms of one-piece base. Head forms are more varied in what was presumably bone bait hooks. No stone sinkers were identified. a single, brief occupation layer in Area D; hooks from Fig.32. (below) Fishing gear from Area AB. A layer 0; Area AB and Area C have only one of the head forms C–G layer 2; B, H layer 3; K layer 6; L, M layer 7; N represented in Area D. The distribution of these items is layer 8; I, J layer 9. shown in Table 2. Archaeological Excavations at Sarah’s Gully 115

Fig.33. Fishing gear from Area C. D, E layer 4; A, C layer 5; B fire feature sealed by Layer 3.

Fig. 34. Fishing gear from the coastal part of Area D. A–D surface collection; H, I layer 4; G inland, baulk 2–3 layer 3; remainder uncertain. 116 Tuhinga, Number 29 (2018)

Table 2 Distribution of fishhooks and related items at Trolling is indicated at the Settlement by what is Sarah’s Gully Settlement. presumably a broken uniperforate lure point from layer 9 Area Layer Hook or Tab or Drill File in Area A (Fig. 32J), and an unbarbed, unperforated point hook part unfinished point from the main cultural layer (4) in Area D (Golson 1959a: AB Surface and 1 9 2 3 – 14), which is no longer present in the collection. A complete bone trolling lure shank and part of a similar 2 6 6 1 2 item were found in the lowest cultural layer of Sarah’s 3 2 6 2 1 Midden. Both are flat and oval-sectioned with dorsoventral 4 1 – 1 – perforation. The complete example has disappeared and is 5 – – 1 – recorded only in a photograph (Fig. 35A); the fragment is 6 1 2 1 2 a snapped-off proximal end, also broken at the perforation (Fig. 35B). The complete example had side notches to 7 2 – – 1 secure the point lashing. A worked piece of silicified wood 8 3 – 1 – from the surface of Area AB is probably a blank for a trolling 9 2 – – – lure shank (Appendix 4 and Fig. A4.1). A similar but only Subtotal 26 16 10 6 slightly worked piece came from layer 2 in the Pä. Only three items of fishing gear were found at the Pä: Area C 3 – – 2 1 two stone sinkers from layer 2 in squares 16 and 16b, and 4 2 1 – – part of a barbed point of mammal bone in an apparently 5 3 – – – deliberately filled posthole in a small rectangular pit in the Subtotal 5 1 2 1 same area as the sinkers. The base of the point is broken off, Area D Surface 3 4 1 2 so the nature of attachment to the shank is not known (Fig. 36A). One of the sinkers weighs 134 g (Fig. 36C). The other Localised, 9 – – – is largely unmodified, but has faint traces of working around no layer its edges. It weighs 76 g and is of a size and shape suitable for 4 5 – – – enclosing in the edge of a net (Fig. 36B). Inland – – – 3 Subtotal 17 4 1 5 Total 48 21 13 13

A different, later style is represented by the point of a two-piece fishhook (Fig. 32A) from the surface of Area AB. It is worth emphasising, however, that the typical ‘early’ one-piece bone hook is present in almost all contexts in the Settlement. In striking contrast, the only one-piece fishhook at Sarah’s Midden is a small point-leg fragment with bait notch from layer 2 of square 102, presumably displaced from its original context. There is no evidence of hook manufacture apart from a fragment of a stone file from layer 5, which could have been used for purposes other than making fishhooks. A different kind of hook is suggested by a small, square- sectioned copper nail, the lower third bent at a right angle to the rest of the nail, from layer 3 in square 102. A small, Fig. 35. Bone lure shanks from flattish oval pebble, slightly notched at each end, from layer 4 Sarah’s Flat. A probably from layer of square 101 appears to be a sinker, perhaps for a small baited 5 (from a photograph by Anne hook, or for a net. Leahy); B layer 5. Archaeological Excavations at Sarah’s Gully 117

were broken is not known, but ritual shattering cannot be ruled out. There are also a number of chips, some quite large, with a high degree of polish, from the deposits in all three areas. One of these, from an uncertain context in Area D, is almost certainly from an adze made from the black-veined, light grey metasomatised argillite best known at Ohana on D’Urville Island, but also found in other places in the Nelson mineral belt. A piece from square F8, layer 2, in Area AB has also been identified by P.R. Moore (pers. comm., 1 June 2016) as ‘dark grey meta-argillite: Nelson–Marlborough’. The other items from the coastal part of the Settlement are mostly broken, unfinished preforms of small to medium size, although there is one butt from a large preform. Most Fig. 36. Fishing gear from the Pä. of these are illustrated in Figs 38–40. The exceptions are a small, partly reworked piece from layer 7 in Area AB; a Adzes and adze working scrappy quadrangular fragment from layer 5 in Area C; and a quadrangular-sectioned butt and a rough preform from layer Most of the stone adze blades and adze-related material from 4 in Area D. Chips from finished adzes were found in all the Sarah’s Gully sites can only be described as small and layers except layer 6 in Area AB, and layers 2 and 3 in Area scruffy. However, a few fragments show that the occupants of D. The latter absence may well be due to poor retention the Settlement, at least, had access to large, well-finished tools, strategy. It seems that people were using and reworking which were not left behind when the people moved on. This existing adzes as well as working small to medium preforms, is best illustrated by the butt of a very large, well-polished at least, at all stages of occupation. It is notable that only one hog-back blade, of which four joining pieces were found small rough adze was found in Area C (Fig. 39A), although in adjacent squares of layer 4, Area D, along with a single a concentration of basalt waste in layer 2 suggests adze mid-section of a well-polished quadrangular-sectioned adze working in that area. Again, the small number from Area D in a darker stone (Fig. 37). How or why these beautiful tools may reflect collection strategy.

Fig. 37. Broken pieces of large well-finished adzes from Area D, layer 4. 118 Tuhinga, Number 29 (2018)

Fig. 38. Adze preforms and fragments from Area AB. A–C layer 1; D–E layer 2; F layer 4; G–H layer 6; I layer 7; J–K layer 9.

Fig. 39. Adzes and performs from Areas C and D. A Area C layer 4; B–D Area D just inland from the coastal excavation. B layer 3; C–D layer 2.

Fig. 40. Adze and performs from Area D Inland. A upper terrace Layer A; B–E lowest terrace layer 3. Archaeological Excavations at Sarah’s Gully 119

Fig. 41. Adze fragments and preforms from Sarah’s Flat. All from layer 5 except C, which is from layer 3.

Adzes and other stone tools were also found in the small hog-backs, while the other two could be either hog- inland parts of Area D. Three items found just inland from backs or reversed-triangular adzes. All would have been the beachfront excavation on the terrace beside the stream medium or small tools when finished. are illustrated in Fig. 39 (B–D). One from further inland on Most of the adzes and preforms from the Pä are illustrated the bottom terrace, and one from the uppermost flat, beside in Fig. 42. The only complete adze – a substantial blade Feature 37, are shown in Fig. 40. Also from this part of the of irregular shape and rounded cross section – has a well- site were found bevel ends of a quadrangular-sectioned adze sharpened, double-bevelled cutting edge (Fig. 42A). Other and one with a triangular section that appears to have been items include the butt of what was probably a very small intended as a reversed-triangular adze (Duff type 3) (Duff hog-back adze (Fig. 42E), and three fragments of what 1956: 171, fig. 38). Both are poorly made. These adzes were probably triangular-sectioned preforms (Fig. 42D and from the inland part of the settlement are indistinguishable F) Two of these (one not illustrated) were from under the from those from the beachfront deposits. defensive bank in square 13A. Some chips from finished Six of the seven broken preforms from Sarah’s Midden adzes were found: six from layer 2, three from layer 3 and are illustrated in Fig. 41. Two were clearly intended to be two from pit fills.

Fig. 42. Adzes and adze fragments from the Pä. A pit fill; B, D–F, J–L layer 2; C posthole fill; G–I layer 3. 120 Tuhinga, Number 29 (2018)

Table 3 Distribution of basalt flakes at the worked by similar techniques into similar forms throughout Sarah’s Gully sites. the occupation sequence. Golson’s comment about the sequence at Pig Bay on Motutapu Island – ‘the remarkable Site and layer No. Weight (g) thing is that the material is Archaic throughout’ (1959a: 46) Settlement – could also be made about the adze material from Sarah’s Area AB, 0–3 615 4308 Gully. Area AB, 4–10 532 3092 Included in the material from Golson’s excavations at Sarah’s Gully, but undocumented, is an unfinished and Total 1147 7400 unusually small example of a ‘side-hafted’ adze (Duff’s type Area C, 3–3A 105 1803 V) (Duff 1956: 191, fig. 48), apparently fashioned from a Area C, 4 16 315 larger adze, as a partly ground surface is more compatible Area C, 5 8 114 with a finished adze than with the preform as it exists now Total 129 2232 (Fig. 43). It could be a surface find from anywhere in the Area D, vicinity, but is important in adding to the known distribution 121 2192 surface and 2 of this rare style. Area D, 3+ 2 301 Area D, 4 102 1984 Total 225 4477 Sarah’s Midden 2 3 36 3 6 175 4 20 221 5 6 157 Total 35 589 Sarah’s Pä 2 73 1795 3 15 293 Features 6 166 Total 94 2274

Adze manufacture and repair was evidently an ongoing task. The distribution of basalt flakes and chips in the sites is shown in Table 3. There were notable concentrations in several places, including layer 7 in Area AB, layer 3 in Area C and layer 4 in Area D. These concentrations included some very large flakes, some with cortex. These suggest that larger preforms may have been manufactured at these locations than the discards actually recovered suggest. Although most of the basalt appears to be waste, there are several blade-like pieces that may be tools in their own right, and at least one possible awl or drill tip. There is no evidence of chronological change in the Fig. 43. Small preform, apparently of a side- adze material from the three excavations. It seems that basalt hafted adze, reworked from a larger tool. Context from the nearby source at Tahanga hill continued to be unknown. Archaeological Excavations at Sarah’s Gully 121

Drill points

The majority of items from the Settlement identified as drill points came from Area AB. Their distribution is reasonably consistent with that of drilled tabs from fishhook manufacture (Table 2), although there may have been many more drills and tabs from Area D that were not included in the official collection from that area. Three broken drill points are from layer 8 of Area AB, and layers 2 and 4 of Area D. A less certain example is from layer 3 of Area AB. With one exception, the points are made from siliceous stone in a wide range of colours, from white to grey, brown and dark red. Most are illustrated in Fig. 44. The exception is a basalt point, now broken, from Area AB, which is the only item in the group not definitely a drill tip. No drill points were found at Sarah’s Midden or the Pä, which is consistent with the lack of evidence of fishhook manufacture in those sites. Fig. 44. Stone drill points from Area AB of the Settlement. D, G, J Layer 1; I Layer 2; A, C, E Layer 3; H Layer 5; B; Layer 6; F Layer 8. Abrading tools No large grinding stones or pieces were found in the Settlement. A broken rectangular stone of fine-grained rock, which had clearly been used as a smoothing stone, is from layer 1 in square 5A on the lowest terrace of the inland part of Area D. It is very regular in shape, 57 cm wide and 38 cm thick, with a gently rounded end and edges. The existing length is 10 cm. It is similar in size and proportions to a stone found on the Pä, but that has no evidence of smoothing (see below). Three burnt pieces of possible grindstones came from squares 11f and 12n in the upper part of Area D inland. A piece of pumice shaped like a sandstone file and possibly used as an abrader was found in layer 5 of Area C. Sandstone files were scattered throughout the Settlement deposits, but were not numerous. Most appear to be broken, which is probably why they were discarded. Their distribution in the beachfront areas is shown in Table 2; some are illustrated in Fig. 45 (F–I). There were also three from the inland part of Area D: squares 2–3, layer 3; squares 6–7, layer 1; and an uncertain context further up. These last were not associated with fishhook manufacture. The only abrading tools from Sarah’s Midden are a small file, tapering at one end (Fig. 45J); a piece of pumice with Fig. 45: Stone files from the three sites. A–E the Pä: A one flat face bearing signs of use as a smoothing tool (both layer 1; B–D layer 2; E layer 3. F–I the Settlement: F Area D layer 3; G Area D layer 2; H Area AB layer 6; I Area from square 101, layer 5); and a tiny fragment of a grindstone AB layer 7; J Sarah’s Flat layer 5. from square 101, layer 3. 122 Tuhinga, Number 29 (2018)

The Pä yielded part of a grindstone and several files. The The sources of the obsidian have not been established. former was found in the fill of a pit; it weighs 1.5 kg and has Sarah’s Gully is close to a number of known sources and it is two dished surfaces. Several files are illustrated (Fig. 45A–E). likely that the people who periodically occupied the various Two other probable abraders – two matching pieces of an sites had access to most or all of them. It would, however, be oval-sectioned piece of sandstone with little or no sign of use, very useful to determine whether they were also obtaining and a broken central section of a larger piece – were found obsidian from more distant sources in Northland and Taupo. together in a context not properly documented. Another The obsidian is all of high quality, without spherulites. part of a probable file came from layer 2 in square 17a. A now The chert has also been described as sinter or siliceous broken pumice pebble measuring 62 mm long, from layer 2, sinter. This term was introduced to the archaeological has clear signs of abrasion in the form of a shallow groove in literature by Best (1975) and Best and Merchant (1976), one surface. An unusual stone with several grooves suggesting who suggested that drill points found at the early use as an abrader (from square 17a, layer 2) was identified as archaeological site of Houhora in the far north of the North grey perlite (perlitic rhyolite). This material is not known on Island were made of sinter from the Coromandel Peninsula. the Coromandel Peninsula north of Whitianga, although it The Kuaotunu Peninsula is home to an extensive area of may occur on Great Mercury (Ahuahu) and Ohinau Islands siliceous sinter (Hamilton et al. 2016). Newton studied what and is common on Great Barrier Island (Aotea Island) (P.R. he called Waitaia sinter, and provided a detailed map of Moore, pers. comm., 1 June 2016). its distribution and variety (Newton 2000: 3). This shows its ready accessibility on the headland and foreshore at the western headland of Otama Beach. It is likely that all the Hammer stones Sarah’s Gully chert is local. Only one definite hammer stone was found, despite the The most common material in the collection is a fine- numbers of basalt flakes and broken preforms. This is a heavy grained, light brown chert, which produces high-quality spherical object with extensive pitting on one flattened conchoidal fracture. This must be abundant in the vicinity surface. It came from square 3, layer 4, in Area D, just inland because the flakes and cores in the assemblage are larger than from the coastal excavation. A pebble suitable for fine those of the other types of material. There is also quite a lot knapping came from layer 3 in Area C of the Settlement. of a fine-grained, very dark to black chert with good-quality Some other apparently unworked pebbles may also have conchoidal fracture. A large core of the light brown chert been brought to the sites for use as hammers. This general shows a gradation to this very dark grey material; they are absence of hammer stones was also noted at Golson’s Pig Bay clearly the same. site, where adze manufacture was a major activity (Davidson The second-most abundant material is chalcedony of & Leach 2017: 30). various types, some pieces of which are also quite large. Most are translucent, and some are transparent and glass- like. They are of various shades, including light pink. Next Other worked stone most abundant is a course-grained chert of various colours. In addition to Tahanga basalt, the people at Sarah’s Gully Some bright red flakes are probably jasper. Finally, there is collected obsidian and a variety of other types of rock for quite a lot of quartz in the collection. fashioning usable hand tools from small pieces. Items of obsidian and what was described at the time of excavation as Mass distribution chert were abundant in the excavations, although distributed somewhat differently. Eight hundred and twenty-one All pieces of obsidian and chert were inspected and weighed pieces of obsidian were recovered, of which 683 were from individually to the nearest milligram to obtain an idea of the the excavations and the remainder from the large surface general size of the pieces being used and discarded. collection in Area D coastal. The chert assemblage consists In any collection of flake debitage, a strong preponderance of 402 items, only 11 of which are from the Area D surface of small flakes and a decreasing number of larger specimens collection. There is more chert from all layers in Area AB would be expected. Consequently, normal size-frequency and layer 4 in Area D, similar amounts of chert and obsidian diagrams are so heavily skewed to the left that comparison from Area C and Sarah’s Midden, and far more obsidian than between assemblages is difficult. One common way to chert from the Pä. observe any central tendency in choice of size is to use a Archaeological Excavations at Sarah’s Gully 123

logarithmic scale. This approach is taken here. To illustrate A suitable nearby assemblage for comparison in Seelenfreund- the point, the two types of distribution curve are shown in Hirsch’s study is that from the Kauri Point Swamp on the Fig. 46. coast of the Harbour (Table 4). The log-transformed In the lower graph, the raw data are used. The right-hand mass data are shown in Fig. 47. There are several notable tail actually extends out to 68.51 g, but is here truncated features in this comparison. The surface-collected specimens at 15 g, otherwise no structure would be observable. In from Sarah’s Gully show a greater preponderance of larger the upper graph, the raw data have been log-transformed, pieces than the excavated material. This is not surprising, using milligrams rather than grams to avoid negative values. as very small items are often ignored in surface collecting. This clearly shows a central tendency, depicting human The Kauri Point collection has far more very small flakes selectivity by size. The excavated sample of 683 pieces than the excavated collection from Sarah’s Gully, which may ranged in size from 78 mg to 68.51 g, and weighed a total reflect a similar tendency not to retain the smallest excavated of 2.646 kg. The arithmetic mean size is 3.87 g. specimens from Sarah’s Gully. Although some sieves were Fortunately, a large body of data is available with used in the excavation, the extent of such use is not known. which the Sarah’s Gully collection can be compared. The analysis of chert mass is summarised in Table 5. Seelenfreund-Hirsch (1985) studied collections from As with the obsidian, a normal size-frequency distribution numerous archaeological sites around New Zealand, would show very little, so the data were log-transformed totalling 14,355 pieces, and documented flake mass and (Fig. 48). It is not surprising that the chert being used was, maximum dimension. on the whole, larger than the obsidian.

Fig. 46. The Sarah's Gully obsidian showing the two methods of depicting the size frequency distribution. 124 Tuhinga, Number 29 (2018)

Table 4 Kauri Point Swamp and Sarah’s Gully obsidian Of the 683 excavated pieces examined, a surprisingly large dispersion statistics. number showed evidence of use-wear, almost all of which was micro-flaking along one or more edges. Since multiple Kauri Sarah’s Settlement, items were originally stored in single plastic bags, some Point Gully Area D, Swamp excavations surface use-wear can be attributed to post-excavation damage. To prevent this happening in future, all items were rebagged in No. 5733 683 124 individual plastic bags and given unique catalogue numbers. Total weight (kg) 21.951 2.646 0.698 Because of the possibility of post-excavation wear and tear, Lightest (g) 0.020 0.078 0.327 only items that showed the most certain pre-European use- Heaviest (g) 135.240 68.51 40.5 wear were selected for description. This amounted to 147 Mean weight (g) 3.829 3.87 5.63 pieces, or about 22% of the collection (the real percentage is probably somewhat higher than this figure). Although any one piece of obsidian may have several Obsidian use-wear patterns edges that might be utilised as tools, by far the most common form of use-wear was seen on only one edge. An edge used Each obsidian item from the excavations at Sarah’s Gully in a sawing action (knife) produces scratches that are parallel was examined with oblique light using a Kyowa SDZ stereo to the edge, while a scraping action produces micro-flakes binocular microscope with 10× eyepiece, 0.5 lens with 0.7– along the edge. If the scraping action is in only one direction, 4.5 zoom and a MEM 1300 digital USB microscope camera. then the micro-flakes occur only on the side opposite the Additional photographs were taken with a Canon EOS 6D direction of the scraping. If a scraper is used back and forth, camera with Macro lens and one barrel extension. then micro-flakes form on both sides of the edge.

Fig. 47: Comparison of the Sarah's Gully obsidian with the Kauri Point Swamp assemblage and the surface collection from Sarah's Gully, using log-transformed mass. Archaeological Excavations at Sarah’s Gully 125

Table 5 Mass analysis of chert. Table 6 Distribution of obsidian tools in the Sarah’s Gully sites. Tool types: A, straight unidirectional scraper; Normal data Log-transformed data B, straight bidirectional scraper; C, concave unidirectional scraper/spokeshave; D, concave bidirectional scraper/ No. 402 – spokeshave; E, convex unidirectional scraper (nose Lightest 129.000 mg 4.860 ln(mg) scraper); F, convex bidirectional scraper (nose scraper); Heaviest 947,000.000 mg 13.761 ln(mg) G, awl; H, knife edge; I, multi-purpose tool. Total weight 5,681,022.000 mg – Mean weight 14,131.896 mg – The Settlement Consequently, such use-wear patterns allow a simple Sarah’s Sarah’s Tool Päā Midden Total classification of tool types: knife and scraper. Scrapers can be Area AB Area C Area D classified as unidirectional or bidirectional. In addition, the A 18 6 7 shape of the edge can be used to subdivide scrapers as straight, B 2 1 2 8 – 13 concave (spokeshave) or convex (nose scraper). A special C 4 2 2 1 – 9 form of scraper is one that shows evidence of rotational use (awl), and the micro-flaking can easily determine whether D – – – 1 – 1 one was rotated clockwise, anticlockwise or in both E 2 – – 10 1 13 directions. F 2 – – – – 2 Finally, a few obsidian pieces had evidence of use on G – – – 5 – 5 several edges; these are classified as multi-purpose tools. H 1 – – 2 – 3 The distribution of these various kinds of tools is shown in Table 6. I 2 – 14 2 18 By far the commonest form of tool was a straight Total 31 9 11 85 11 147 unidirectional scraper (56%). Some typical examples of the tools found in the collection are described below.

Fig. 48. The Sarah's Gully chert size-frequency distribution, using log-transferred mass. 126 Tuhinga, Number 29 (2018)

Awl-shaped tools Three awl-shaped objects from the Pä are illustrated in Fig. 49. Each has a triangular cross section, and each has been used in a distinctly different manner. SG758 has unidirectional micro-flaking in opposite directions on each side, indicating rotational use in only one direction, clockwise. This qualifies as an awl. SG731, on the other hand, has micro- flaking in only one direction on both edges and cannot have been used in a rotational action. Instead, this object must have been used as a unidirectional scraper to make a trough. There is a third unidirectional scraper edge on this object. SG479 has clear bidirectional micro-flaking on all three edges, indicating its use as an awl by rotating in both directions.

Fig. 49. Awl-shaped tools from Sarah's Pä. SG758, SG731: layer 3, SG479: layer 2.

Concave scrapers/spokeshaves Two concave scrapers/spokeshaves from the Settlement appear in Fig. 50. In both cases, the micro-flaking is unidirectional. Note the heavy striations on SG134. These are natural ripples in the glass surface that sometimes appear during conchoidal fracture and might easily be misinterpreted as use-wear.

Fig. 50: Concave scrapers/spoke-shaves from the Settlement. SG6: Area AB layer 1, SG134: Area D coastal layer 3+

Knives Only three items found could reliably be identified as having been used as knives. One from Sarah’s Pä is illustrated in Fig. 51. It has minute scratches, horizontal to the cutting edge, and some micro-flaking along 32 mm of the edge. In the enlargement, the scratches are easily visible.

Fig. 51: Knife from Sarah's Pä, layer 3. Archaeological Excavations at Sarah’s Gully 127

Straight-edged unidirectional scrapers Five of the 83 scrapers in this category are illustrated in Fig. 52. Although, strictly speaking, the edges of these are not always straight, those that are concave or convex are much more obviously chosen by the stone workers because of their non-straight-shaped edge. The five scrapers are: SG3, 14 mm wide, slight concavity; SG37, 21 mm long, slightly convex edge; SG132, 29 mm; SG386, very heavy flake damage along 24 mm; SG429, micro-flaking along 48 mm of slightly convex edge.

Fig. 52: Straight-edged uni-directional scrapers. SG3: Area AB layer 1; SG37: Area AB layer 3; SG132: Area D coastal layer 3+; SG386: Sarah's Pä layer 1; SG429: Sarah's Pä layer 2.

Straight-edged bidirectional scrapers One of these scrapers, from Area C of the Settlement, is illustrated in Fig. 53. It is a small block of obsidian. The vertical right-hand edge in the image has bidirectional use-wear flaking along 22 mm. The upper horizontal edge in the image shows considerable attempts to remove small flakes. This is not use-wear.

Fig. 53: Straight-edged bi-directional scraper (right hand vertical edge). Area C: layer 5. 128 Tuhinga, Number 29 (2018)

Convex unidirectional scrapers (nose scrapers) Ten of these tools were identified. An example from Sarah’s Pä is illustrated in Fig. 54. The use-wear extends 31 mm along each side of the ‘nose’. Although this item has an awl shape, the term awl is normally reserved for a tool that is used for drilling holes, in which case the micro-flake use-wear is either on alternate edges (rotating in one direction), or on both edges (rotating back and forth in both directions). An awl-shaped tool that has micro-flake use- wear on both edges in the same direction must have been used as a scraper, presumably to make a trough-shaped depression in some object. A similar function is interpreted for SG731 in Fig. 49.

Fig. 54: Convex uni-directional scraper/ nose scraper. Sarah's Pä, surface of layer 3.

Fig. 55: Multi-purpose tools. Sarah's Pä, layer 2.

Multi-purpose tools Only about 12% of the tools identified from use-wear had more than one function, in the sense that more than one area of the object displayed signs of use. Two of these, both from Sarah’s Pä, are illustrated in Fig. 55. SG576 has two worked edges, one 21 mm long and the other 14 mm long; both edges are straight and show unidirectional micro-flaking. SG416 has three worked edges, all showing unidirectional micro-flaking; two are straight (24 mm and 13 mm), and one is convex and 37 mm long.

Fine retouching on obsidian by pressure flakingA number of items from the Sarah’s Gully sites display clear evidence of pressure flaking, rather than cruder, hammer-stone knapping. An example from Area C of the Settlement is illustrated in Fig. 56. SG105 is not a recognisable tool (36 mm maximum dimension), but shows very fine, shallow surface flaking. This is reminiscent of advanced flaking technologies like the Solutrean, which use the outre passé, or overshot pressure-flaking technique to reduce the thickness of an object without reducing its width. Intentional thinning does not seem to be the purpose with SG105, or other examples of the same pressure-flaking technique in the sites (see SG129 in Fig. 51 and SG576 in Fig. 55). By way of contrast, the method of flaking by hammer-stone knapping can be observed on item SG728 in Fig. 54. Fig. 56: Fine re-touching on obsidian by pressure flaking. Area C: layer 3. Archaeological Excavations at Sarah’s Gully 129

Chert usage Several varieties of chert were used for the items identified as drill points, thought to have been used in the manufacture of bone fishhooks and described above (Fig. 44). Apart from these items, visible evidence of use-wear on chert flakes is very rare. One item from layer 9 in Area AB appears to be a concave unidirectional scraper. The collection was examined under low-power binocular microscope for signs of retouch and use-wear. Although a few pieces showed secondary chipping, most were undistinguished in this respect, possibly because the bulk of the specimens are fairly coarse-grained rock. Even the best- quality chalcedony showed little evidence that it had been fashioned into usable tools. More than three-quarters of the pieces are best described as broken-down chunks rather than flakes. Two unusual items are illustrated in Fig. 57, along with a similar piece in basalt. These two blade-like flakes Fig. 57. Unusual examples of flake tools from Area AB. are from layers 1 and 3 in Area AB. They were initially The chert items are from layers 1 and 3; the basalt is a identified as silcrete, possibly from the Nenthorn source surface find. in Otago, by Foss Leach and Atholl Anderson. Graeme Mason (pers. comm., 20 April 2013) also identified them as silcrete, but considered them unlike any South Island source he knew. P.R. Moore (pers. comm., 1 June 2016) has subsequently identified them as ‘coarsely textured chert with minor crystalline quartz and chalcedony veinlets’ from the Coromandel, possibly Kuaotunu–Otama. This material is in fact the same as that described above (in the section ‘Other worked stone’) as ‘a course-grained chert of various colours’, and is almost certainly local.

Part of the stem of a clay pipe was found in the surface Miscellaneous sand of Area AB in the line F/H7. It is 55 mm long. It is the only European item found anywhere in the Settlement, An unusual item from the Pä is an adze-shaped piece of although items of European origin were found in the upper coarse-grained stone. It has a slightly rounded rectangular layers of Sarah’s Midden. section and measures 123 x 50 x 29 mm. It appears have A small, damaged, conical-shaped pumice object came been pecked on the sides, but there is no obvious evidence from the inland part of Area D (baulk 12m–n, layer X). The of working or use on either end. It was found in layer 2 in top diameter is about 35 mm and the height 40 mm. It was square 15C, the same context as the butt of a very small probably a gourd stopper. A smaller, less carefully shaped hog-back adze (Fig. 42E). It is very similar in shape to the example with red staining on the upper surface was found in smoothing stone from layer 1 in square 5A of Area D inland, layer 3 in the Pä. There is also a possible worked fragment of described above. pumice from layer 3 in Area AB. 130 Tuhinga, Number 29 (2018)

Other portable items Settlement) is heat-shattered debris from oven stones of local andesite and is not further considered here. There is Pumice also a range of pebbles (smooth round or oval stones) and Unworked pieces of pumice (mostly small) were found in irregularly shaped stones of various kinds. Some of the various contexts in each of the sites. pebbles are suitable for hammers or smoothing stones, and The largest amount is from the Settlement, with 31 may have been brought to the sites for such purposes. pieces from Area AB, from all layers except 5 and 9, and Small, highly polished pebbles, regarded at the time of the 15 from layers 3 and 5 in Area C, of which two from layer excavation as moa crop or gizzard stones, were particularly 3 show possible signs of use as abraders. Only one piece was numerous in the Pä (11 from layer 1, 200 from layer 2, retained from the coastal excavation in Area D (a surface 167 from layer 3, 16 from under the outer defensive bank, find); there are two from square 4 on the lowest inland and 10 from the fills of pits or postholes). There was also a terrace and three (two of which were once part of one large concentration (111) in squares n–p in Area D inland, where piece) from the uppermost excavation. Although pumice there were a further 78 pebbles. These last may represent an from the beach could have blown into the coastal deposits, actual concentration or simply reflect a change in collection it is likely that some, at least, and certainly the large pieces strategy in this particular area. P.R. Moore (pers. comm., 6 from the inland part of Area D, were carried in by people. June 2016) discusses them as follows: Similarly, pumice in the Pä is most likely to have been The fact that they are virtually all composed of the same brought in by people, although as the pieces are all fairly material would suggest they were purposely selected, small, they could have come in with fishing gear or even unless small pebbles on the nearby beach consist almost been brought in by children. There are 10 pieces from layer exclusively of chert with a similar degree of rounding. 2, two from layer 3 and one from a pit fill. Some of the chert is typical of that from the Waitaia The only pumice retained from Sarah’s Midden is from Sinter, in terms of colour, texture and the presence layer 5. It consists of 14 small pieces, which could well have of fossil plant remains. None is obviously exotic been on the sandy surface when the site was first occupied. (i.e. non-Coromandel). Therefore if the stones Ochre represent gastroliths then the animal(s) obtained the A few pieces of crumbly red material were probably bulk of them locally. This would probably rule out intended to be used for kököwai (ochre). There are single seals but not moa. Stones thrown up onto land as a pieces from layers 1 and 4 in Area AB, and from layer 2 of result of major storms or tsunami would be expected to square 2A and layer X of square 12n in Area D inland. There include a wider range of lithologies and perhaps sizes. are also two pieces of similar but yellow material from the Unusual stone included rod-like pieces of silicified wood last context and one from layer 4 in Area AB. Several red from the Pä and Areas AB and D of the Settlement. Two pieces were found in square 18A and one in baulk 16a/17a worked pieces, mentioned above, may have been blanks for in the Pä, all in layer 2. trolling lures. There are three (and probably another one) Kauri gum from layer 2 at the Pä, four from layer 2 of square 4 in Area D The small amount of kauri gum retained is likely to have inland, and one from layer 8 of Area AB. This last is the only derived from the local environment. As might be expected, piece from a definite early context. Silicified or petrified all examples from the Settlement are from Area D inland: wood is found in many parts of the Coromandel Peninsula about five largish pieces (now fragmented) from the natural (Moore & Wallace 2000) and is particularly abundant on surface in squares 15m and 15n; and a largish piece from nearby Great Mercury Island (Ahuahu) (Appendix 4). All square 4–4A and a smaller piece from square 2A, both from these pieces would have been brought to the sites deliberately. layer 3. There are several from the Pä: one from square 16B Two unused, rough, small pieces of quartz, probably and two from square 13A, all from layer 2; and two tiny local, came from the Pä (layers 1 and 2) and one from Area pieces from layer 3 of square 16A. D inland (baulk 4–5, layer 2); there was a piece of grey chert with quartz veins from the Pä (layer 2), a piece of poor- Unworked stone quality chert with quartz veins from layer 2 of Area AB, A substantial amount of non-artefactual stone was retained and an unworked pebble of the similar material from Sarah’s from the excavations. Some of it (nearly all from the Midden (square 100, layer 2). Archaeological Excavations at Sarah’s Gully 131

Table 7 Charcoal identifications from thePä and Sarah’s Midden. L = layer. Pä Sarah’s Midden Species Plant type L2 L3 Pits % L3 L4 L5 % Bracken Fern – 1 – 0 3% – – – (Pteridium esculentum) . Shrub spp. – 3 1 – – – Tutu (Coriaria sp.) 3 2 5 – – – Hebe (Hebe sp.) 17 11 11 1 – – Coprosma (Coprosma sp.) 15 14 21 11 – 8 Mänuka 10 – (Leptospermum scoparium) – – – – Ngaio (Myoporum laetum) 1 11 19 3 – – Mähoe ( ramiflorus) 2 1 2 – – 1 Mingimingi (Leucopogon fasciculatus) – – – 5 – – Pittosporum (Pittosporum sp.) 6 2 2 – – 1 Lancewood (Pseudopanax crassifolius) – – – 10 – – Shrub or scrub Whärangi ( ternata) 1 2 1 – – 1 spp. Olearia (Olearia sp.) 1 – – 83% – – 1 5% Puka (Griselinia lucida) – – 1 – – – – Patë (Schefflera digitata) 1 – – – – – Hangehange – 1 – – – 1 (Geniostoma ligustrifolium) Kawakawa (Piper excelsum) – – 1 – – – Fivefinger – 5 1 – – – (Pseudopanax arboreus) Ribbonwood or – – – – – 3 Plagianthus sp. Mäpou (Myrsine australis) – 2 – – – – Känuka 20 23 28 – – – (Leptospermum ericoides) Rätä vine – – 5 – – – () Tawa (Beilschmiedia tawa) – 1 – – – – Püriri (Vitex lucens) 8 3 14 11 – 11 Broadleaved trees 13% 32% Pöhutukawa – 14 – 11 – 22 (Metrosideros excelsa) Kauri (Agathis australis) 8 1 – – 1 – Kahikatea Conifers – – 1 3% – – – 1% (Dacrycarpus dacrydioides) Mataï (Prumnopitys taxifolia) – – – 1 – – Total 93 98 112 303 53 1 49 103 132 Tuhinga, Number 29 (2018)

Table 8 Charcoal identifications from Areas AB and C at the Settlement. L = layer.

Area AB Area C Species Plant type L1 L2 L3 L6 L7 % L3 L4 L5 Lf % Tutu (Coriaria sp.) – 6 1 – – – – – – Hebe (Hebe sp.) – 8 5 – – – – – – Coprosma (Coprosma sp.) 4 22 6 1 6 – – – – Fivefinger (Pseudopanax arboreus) – 2 – – – – – – – Ngaio – – 3 – 1 – – – – (Myoporum laetum) Olearia (Olearia sp.) – – – – – – – 1 – Pittosporum – 1 – – – – – – – (Pittosporum sp.) Shrub or Akeake – – – – – 52% – – – 2 19% (Dodonaea viscosa) scrub spp.

Mingmingi 2 2 – – – – – – – (Leucopogon fasciculatus) Whärangi (Melicope ternata) – – – – 1 – – – – Mänuka (Leptospermum scoparium) – – 3 – – – – – – Toro (Myrsine salicina) – – – – – – – 10 – Mäpou (Myrsine australis) – 2 – – – – – – – Mähoe (Melicytus ramiflorus) 1 1 – 1 1 – – – – Püriri (Vitex lucens) – – – – – 11 – – 4 Broadleaved Pöhutukawa 76% trees 10 23 17 3 4 7 4 19 8 (Metrosideros excelsa) Kauri (Agathis australis) – – – – – Kahikatea Conifers 6% 2 2 – – 6 – – – 4 (Dacrycarpus dacrydioides) Total 19 69 35 10 25 153 18 4 30 18 70

Charcoal and vegetation The results according to context are shown in Tables 7–9; the totals from the three excavation sites are compared in Charcoal samples from the Sarah’s Gully sites were analysed Table A5.1 in Appendix 5. Most of the identified charcoal by Wallace in 2003 and 2007; the results are summarised and is from contexts that are either undated, or known to be discussed in Appendix 5. He describes the probable forest relatively recent. It thus reflects a landscape considerably composition in the area immediately before occupation and altered from that found by the earliest arrivals, as Wallace finds that, by and large, the charcoal assemblages reflect a concludes. The probable early contexts are layers 6 and 7 landscape from which forest had already been cleared before in Area AB of the Settlement (three samples); layer 5 in occupation. Area C (three samples) and arguably the other five samples Archaeological Excavations at Sarah’s Gully 133

Table 9 Charcoal identifications from Area D inland at the Settlement.

Species Plant type Lowest terrace Old surface Layer X Fire features Pits % Bracken Fern – 6 – – – 3% (Pteridium esculentum) Tutu (Coriaria sp.) – 5 6 8 5 Hebe (Hebe sp.) 2 4 19 11 6 Coprosma (Coprosma sp.) – 13 16 14 8 Olearia (Olearia sp.) – – 1 – – Pittosporum (Pittosporum sp.) 3 – – – –

Whärangi Shrub or (Melicope ternata) – – – 3 – scrub spp. Ribbonwood (Hoheria 80% – – – – 2 or Plagianthus sp.) – Akeake (Dodonaea viscosa) – 2 – 5 –

Ngaio (Myoporum laetum) – 3 1 12 1 Mänuka 2 9 1 7 7 (Leptospermum scoparium) Mähoe (Melicytus ramiflorus) – – 1 – – Taraire – – – 1 – (Beilschmiedia taraire) Broadleaved 10% Pöhutukawa trees – 5 3 11 1 (Metrosideros excelsa) Kauri (Agathis australis) 3 – – – – Kahikatea Conifers 3% 2 – 1 – 1 (Dacrycarpus dacrydioides) Total 12 47 49 72 31 211

from Area C; and layer 5 at Sarah’s Midden (three samples). pit fills on thePä . Wallace explains that pöhutukawa and Even these do not suggest anything like a pristine landscape. pūriri (Vitex lucens) tend to survive on otherwise cleared Unfortunately, no charcoal was retained from layer 4 in the landscapes. However, Beever et al. (1969: 56) suggested that coastal part of Area D. the relatively young open canopy of pöhutukawa on Cuvier There are, however, some interesting points. Pöhutukawa Island (Repanga Island), just north of the Coromandel (Metrosideros excelsa) is the most ubiquitous species across Peninsula, probably developed following eighteenth-century all contexts, absent from only two small samples in Area clearance by Mäori of forest on the island. Similar processes D inland, layer 4 at Sarah’s Midden (which yielded only of rapid recolonisation may have followed Mäori clearance one piece of kauri) and, more surprisingly, layer 2 and the of forest on parts of the Coromandel Peninsula. 134 Tuhinga, Number 29 (2018)

The Pä charcoal is interesting in other ways, too. Känuka Fauna (Kunzia ericoides), which is not represented at all in Sarah’s The faunal assemblages from the three excavation sites have Midden or the Settlement, is the single most numerous had varied post-excavation histories of storage and analysis. species at the Pä, where it even outnumbers the commonest shrubs – coprosma and hebe species – in all contexts. It is Much of the material was taken to Canberra, but some from thus unlikely to relate specifically to any defensive features. Sarah’s Midden and the Pä was retained in Auckland by the One of two charred kauri post fragments came from baulk Birks. 10–10a and is probably from one of the two square postholes Bird remains from the Settlement and Sarah’s Midden on the inner defensive bank; the other, however, was from have previously been analysed and are available for further a small posthole in square 17 on the edge of the main study, as are the mammal remains from the three excavation concentration of features. sites. Small assemblages of fish remains from Sarah’s Midden Wallace compares the Sarah’s Gully charcoal with and the Pä are available for further study, but only a few fish studies from two other sites in the area. T11/914, on the remains from the Settlement appear to have survived. Shells northern outskirts of Whitianga, appears to indicate the from the three excavation sites are also available for further state of the near-pristine forest in the region at the time of study. first settlement – evidence lacking from the Sarah’s Gully Retention methods were not described at the time of samples. This is not surprising, as there is very little charcoal the excavations. One photograph depicts a sieve in use in from the early layers at Sarah’s Gully. The relevant part of Area AB. The mesh size is not apparent but was probably ¼ T11/914 is thought to date to the fifteenth century (New in (6.35 mm). It is not known how extensively sieves were Zealand Archaeological Association 2009). The other site is used. a large ridge settlement (T10/777) at nearby Opito, which was not unlike Area D inland at the Sarah’s Gully Settlement, Birds although much more extensive. Three areas were dated as Bird remains were initially identified by Ron Scarlett, late fifteenth to sixteenth centuries, sixteenth century, and who summarised the results from Sarah’s Gully Settlement post-AD 1650. A number of species present at the Opito site were not found at Sarah’s Gully and vice versa. There was a (including the Cross Creek surface collection) and Opito far greater representation of kauri and less of pöhutukawa at (Scarlett 1979: 77–78). A more detailed listing of these bird T10/777, while känuka seems to have been absent. Wallace identifications was available for the present study (Scarlett suggests that the kauri in T10/777 was from residual logs and n.d.). The bird bones from the Settlement have been stumps surviving in the landscape long after forest clearance. restudied by Trevor Worthy, who also analysed the bird The species’ absence at Sarah’s Gully is further evidence that bones from Sarah’s Midden. His results are included here. much of the charcoal from there is more recent, dating to a Ten fragmentary bird bones, including parts of long time when the residual kauri resource had been exhausted. bones and ribs, were recovered from various contexts in At T10/777 a clear distinction was made by the the Pä. Most were unidentifiable, but there is a petrel bone excavators between charcoal from fire features and that from (not identified to species) from a pit fill and two little blue general occupational debris. The earlier fire features (which penguin (Eudyptula minor) bones from layer 2 in baulk did not include large ovens) were dominated by kauri, with 16a/17a. pöhutukawa, mänuka (Leptospermum scoparium), olearia At Sarah’s Gully, as at Marfell’s Beach (Worthy 1999: and hebe also common. Later, olearia and pöhutukawa 135–137) and some other locations around the New increased significantly and quantities of kauri dropped. At Zealand coast, remains of moa and other birds are present Sarah’s Gully, the distinction between fire features and other in natural sand deposits underlying the cultural layers. domestic deposits is clearest in Area D inland, where there Scarlett excavated the partial remains of such a moa in what is little difference between the two; shrub species dominate each assemblage. This could well indicate that, despite he described as Late Pleistocene sand below the bottom Golson’s stratigraphic argument for the contemporaneity of cultural layer in square G16 in Area B of the Settlement, the inland pits with the coastal midden, the features of Area and noted that Arthur Black had previously recovered bones D inland are relatively recent. of the same bird when he dug into the area, exposing the It is clear that more large charcoal assemblages would large oven in the section above, before Golson’s excavation greatly expand understanding of vegetation history and began. Single bones of a little blue penguin and a black shag Mäori use of timber in this part of the Coromandel Peninsula. (Phalacrocorax carbo) were also recovered from this natural Archaeological Excavations at Sarah’s Gully 135

Table 10 Minimum number of birds identified from Areas A and B of the Settlement. Number of identified specimens in brackets; L = layer; juv. = juvenile; Nat. = natural sand deposits below cultural levels (* indicates remains of a naturally dead moa).

Species L1 L3 L4 L5 L6 L7 L8 L9 Nat. Total Grey duck (Anas superciliosa) – – – – – 1 (1) – – – 1 North Island kökako (Callaeas wilsoni) – 1 (1) – – – – – 1 (3) – 2 Extinct crow (Corvus moriorum) – – – – – 1 (1) – – – 1 Little blue penguin (Eudyptula minor) 2 (6) 1 (1) 1 (1) – – 2 (6) – – 1 (1) 7 Weka (Gallirallus australis) 1 (2) – – – – – – – – 1 Kelp gull (Larus dominicanus) – 1 (1) – – – – – – – 1 Red-billed gull (Larus novaehollandiae) – – – – – 1 (1) – – – 1 Käkä (Nestor meridionalis) – 1 (1) 1 (1) – – – – – – 2 Mantell’s moa (Pachyornis mappini) 1 (1) – – – – – – – 1* 2 White-faced storm petrel 1 (1) – – – – – – – – 1 (Pelagodroma marina) Northern diving petrel – – – – 1 (1) – 1 (1) – – 2 (Pelecanoides urinatrix) Black shag (Phalacrocorax carbo) – – – 1 (1) – – – 1 (1) – 2 Shy (?) (Thalassarche cf. cauta) – – – – – – – – 1 (1) 1 Total 5 4 2 1 1 5 1 2 3 24

Table 11 Minimum number of birds from Area C of the Settlement. Number of identified specimens in brackets; juv. = juvenile.

Species Layer 1 Layer 3 Layer 4 Layer 5 Features Total ? North Island kökako (Callaeas wilsoni) – – – – 1 (1) 1 Parakeet, käkäriki (Cyanoramphus sp.) – – – 1 (2) – 1 Little blue penguin (Eudyptula minor) 1 (1) 2 (4) 2 (7) 4 (13) 1 juv. 1 (1) 10 Weka (Gallirallus australis) – – 1 (1) juv. – – 1 Kelp gull (Larus novaehollandiae) – 1 (1) – 1 (1) – 2 Käkä (Nestor meridionalis) – 1 (1) – – – 1 Black shag (Phalacrocorax carbo) – – 1 (1) – – 1 Shy albatross (?) (Thalassarche cf. cauta) – 1 (1) 1 (1) – – 2 Total 1 5 5 6 2 19 136 Tuhinga, Number 29 (2018)

Table 12 Minimum numbers of individuals of birds from Area D of the Settlement. Number of identified specimens in brackets; juv. = juvenile.

Species Surface and layer 2 Layer 3 Layer 4 Total Grey duck (Anas superciliosa) 1 (2) – – 1 North Island kökako (Callaeas wilsoni) – 1 (1) – 1 North Island giant moa ( giganteus) – – 1 (1) 1 Little blue penguin (Eudyptula minor) 4 + 1 juv. (13) – 2 (8) 7 New Zealand pigeon, kererü (Hemiphaga novaeseelandiae) – – 1 (1) 1 Kelp gull (Larus dominicanus) 1 (1) – 1 (1) juv. 2 Red-billed gull (Larus novaehollandiae) 1 juv. (1) – – 1 Käkä (Nestor meridionalis) 1 (1) – – 1 Saddleback, tiëke (Philesturnus rufusator) – 1 (1) – 1 Flesh-footed shearwater (Puffinus carneipes) 1 (1) – – 1 Fluttering shearwater (Puffinus gavia) 1 (2) – – 1 Sooty shearwater (Puffinus griseus) 1 (1) – – 1 Shy albatross (?) (Thalassarche cf. cauta) 1 (2) – – 1 Total 13 2 5 20

Table 13 Minimum numbers of birds from Sarah’s Midden. Number of identified specimens in brackets.

Species Layer 2 Layer 3 Layer 5 North Island brown kiwi (Apteryx mantelli) – – 1 (1) North Island adzebill (Aptornis otidiformis) – – 1 (1) North Island kökako (Callaeas wilsoni) – – 1 (2) Parakeet, käkäriki (Cyanoramphus sp.) – – 2 (4) Slender bush moa (Dinornis struthoides) – – 1 (2) Little blue penguin (Eudyptula minor) – 1 (1) 3 (19) Weka (Gallirallus australis) – – 1 Kelp gull (Larus dominicanus) 1 (1) – – Käkä (Nestor meridionalis) – – 1 (2) Fairy prion (Pachyptila turtur) – – 1 (1) Pied shag (Phalacrocorax varius) – – 1 (2) Tüï (Prosthemadera novaeseelandiae) – – 1 (1) Little shearwater (Puffinus assimilis) – – 1 (1) Flesh-footed shearwater (Puffinus carneipes) – – 1 (1) Fluttering shearwater (Puffinus gavia) – 1 (1) 1 (1) Sooty shearwater (Puffinus griseus) – 1 (1) – Total 1 3 17 Archaeological Excavations at Sarah’s Gully 137

deposit. Scarlett identified the moa as Euryapteryx exilis, Similarly, a number of other pieces of moa bone, now considered a of the coastal moa (E. curtus) undetermined to species, were found in layer 5 of Sarah’s (Worthy & Holdaway 2002: 74, 78). He recovered mainly Midden, although Worthy suggests that two might be from ribs and vertebrae, as well as a partial right tibiotarsus and a small emeid moa (i.e. not Dinornis). the left ischium, and commented: ‘This bird had apparently In view of the evidence of fishhook manufacture from drifted for several yards after decomposition, down the old moa bone in all three areas of the Settlement, much of beach, and was very soft and friable (very damp) when I this material is probably industrial and does not necessarily excavated it, after the Moahunter layer above had been represent moa killed or eaten at or near the site. However, finished, but after lying in paper bags in the Anthropology the presence in Area D of tracheal rings, pelvic and rib Dept. store room for a year or so, had hardened considerably’ fragments, and rat-gnawed bones suggests that freshly killed (Scarlett n.d.: 11). Worthy considers this bird to be Mantell’s moa carcasses or parts thereof were brought to this part of moa (Pachyornis geranoides, formerly Pachyornis mappini) the site, at least. The best evidence for actual consumption (see Gill et al. 2010). His opinion is based on the fragments of moa comes from Sarah’s Midden, where a left fibula of the right tibiotarsus. and partial left tibiotarsus of a slender bush moa (Dinornis Minimum numbers of individuals (MNI) and number of struthoides) appeared to be in clear association with a identified specimens (NISP) of birds from each of the three hängï, whereas there was no clear evidence of fishhook beach-front excavation areas of the Settlement and Sarah’s manufacture. Midden are given in Tables 10–13. The identified birds are summarised according to habitat Although identified moa are shown in only three in Table 14. Although some of the terrestrial birds clearly contexts at the Settlement (Area AB, layer 1 and natural; preferred a forest habitat, others are more difficult to classify Area D, layer 4), other fragments of moa bone, sometimes in this way, some preferring more open country and forest worked, were found throughout these deposits as follows, fringe, and some specifically coastally oriented. with Scarlett’s tentative identifications where given: Worthy and Holdaway (2002) gave considerable Area AB attention to the habitats of moa and other now extinct Layers 2, 3 and 3A ? Slender bush moa (Dinornis birds. They noted that the North Island giant moa (Dinornis struthoides); drain fill below layer 3 giganteus, now D. novaezelandiae), the slender bush moa, Layers 4, 5, 7, 9 Fragments too small for identification. Mantell’s moa and the coastal moa were all found in natural deposits at Waikuku and Tokerau Beaches in Northland Area C (ibid.: 197). This is a similar environment to Sarah’s Gully. Layer 3 One fragment Dinornis sp., other fragments too They suggest that the extinct crow (Corvus moriorum) seems small for identification (one burnt). to have been found mainly in coastal areas, perhaps preying Layer 3A One cut fragment. on penguin and seal colonies or feeding in estuaries (ibid: Layer 4 Fragments. 439). The extinct adzebills (Aptornis spp.) are thought to have Layer 5 Burnt fragment. been more prevalent in drier East Coast regions and therefore to have preferred shrublands and grasslands (ibid: 402). Area D

Layer 1 Cut fragment from medium-sized moa. Layer 2 Fragment. Layer 3+ Pelvic fragment and rat-gnawed sacral rib fragment.

Layer 4 Dinornis ?struthoides, numerous fragments of tibiotarsus, tracheal ring. Uncertain context Fragments, probably Dinornis, tracheal ring; Dinornis ?struthoides, sternal rib, rat-gnawed fragment. 138 Tuhinga, Number 29 (2018)

Table 14 Minimum number of birds from Sarah’s Midden (SM) and the Settlement (Areas AB, C and D) according to habitat.

Habitat Species SM AB C D Total Forest North Island kökako (Callaeas wilsoni) 1 2 ?1 1 5 Parakeet, käkäriki (Cyanoramphus sp.) 2 – 1 – 3 New Zealand pigeon, kererü (Hemiphaga novaeseelandiae) – – – 1 1 Käkä (Nestor meridionalis) 1 2 – 1 4 Saddleback, tïeke (Philesturnus rufusater) – – – 1 1 Tüïū(Prosthemadera novaeseelandiae) 1 – – – 1 Subtotal 5 4 2 4 15 Forest fringe/ Weka (Gallirallus australis) 1 1 1 – 3 scrub North Island adzebill (Aptornis otidiformis) 1 – – – 1

Forest/coastal Extinct crow (Corvus moriorum) 1 – – 1 North Island brown kiwi (Apteryx mantelli) 1 – – – 1 Coastal/forest Moa (Dinornithiformes) 1 2 – 1 4 fringe Coastal/inlets Grey duck (Anas superciliosa) – 1 – 1 2 Subtotal 4 5 1 2 12 Little blue penguin (Eudyptula minor) 4 7 10 7 28 Black-backed gull (Larus dominicanus) 1 1 2 2 6 Red-billed gull (Larus novaehollandiae) – 3 1 1 5 Fairy prion (Pachyptila turtur) 1 – – – 1 White-faced storm petrel (Pelagodroma marina) – 1 – – 1 Diving petrel (Pelecanoides urinatrix) – 2 – – 2 Flesh-footed shearwater (Puffinus carneipes) 1 – – 1 2 Allied shearwater (Puffinus assimilis) 1 – – – 1 Fluttering shearwater (Puffinus gavia) 2 – – 1 3 Sooty shearwater (Puffinus griseus) 1 – – 1 2 Black shag (Phalacrocorax carbo) – 2 1 – 3 Pied shag (Phalacrocorax varius) 1 – – – 1 (Shy) albatross (?) (Thalassarche cf. cauta) – 1 2 1 4 Subtotal 12 17 16 14 59 Total 21 26 19 20 86 Archaeological Excavations at Sarah’s Gully 139

At each of the excavated areas at Sarah’s Gully Settlement, scapula (1), mandible (1), premaxilla (1) and syn (1). The the number of seabird remains is more than double that of proportions of penguin bones from Sarah’s Midden, all from forest, forest fringe, scrub and coastal (but not marine) birds. one square, were slightly different, comprising 6 tibiotarsi, One of the seabirds, the black shag, nests in trees and ledges 4 coracoids, 3 femora, 3 humeri, 1 tarsometatarsus, 1 radius up river valleys, as well as on coastal cliffs and rocks. Some and 1 anterior sternum. of the seabird species recorded now nest mainly on offshore No bone bird-spear points were found in the excavations. islands, but at the time of early human occupation of the Wooden spear tips might have been used, but the forest Kuaotunu Peninsula, many may have been breeding there. birds may have been captured in snares, and some of the Although only a few juvenile birds have been identified, flightless birds hunted with dogs. It is possible that some this may be partly due to retention practices. There is no evidence to indicate systematic muttonbirding, as suggested or even most of the seabirds were captured in fishing nets. at Parker’s Midden at Opito (Davidson 1979: 188). Certainly, it appears that a wide range of birds of varying A suggestion of seasonal human occupation is supported habitats was available in the vicinity of Sarah’s Gully during by the presence of flesh-footed shearwater (Puffinus the early occupations. carneipes) in Area AB and white-faced storm petrel (Pelagodroma marina) in Area D. These two species are Mammals not normally present in New Zealand during the winter Scarlett initially provided mammal identifications for (Falla et al. 1970: 41, 56). Their presence therefore suggests the Settlement, which were entered in the Sarah’s Gully summer occupation by humans, but their absence (and from catalogues. He compiled a separate list of dog remains. The such small samples) does not necessarily rule out winter dog remains were also studied by a Taihape vet, G.C. Herd, occupation. The presence of a few juvenile birds (penguins, whose unpublished report is included here (Appendix 6) as gulls and a weka, Gallirallus australis) also suggests summer an interesting early example of archaeozoology, other than occupation. On the other hand, the presence of the shy bird studies, in New Zealand. albatross or shy mollymawk (Thalassarche cauta), if indeed it No records survive of a further study of what were is that species, could indicate winter occupation, as it breeds assumed to be dog coprolites, referred to in Herd’s paper only on the subantarctic islands to the south of New Zealand and listed by context. Ten of the 11 samples Herd examined and Australia. However, these few bones could be industrial, are still present in the collection. These are from layers 6 and rather than food remains (see Rowland 1977: 140, on the 9 of Area AB, layers 4 and 5 of Area C, layers 2, 3+ and 4 Tairua site). Area D, and uncertain. Herd observed splinters of bird or Although relatively few forest bird remains are recorded, rat bone in AB layer 9; largish fish vertebral fragments and they do indicate the presence of some forest habitat in the spicules of bird bone in one of the C layer 5 samples, and vicinity. The kākā (Nestor meridionalis) is the forest bird most bird bone fragments in another; numerous spicules of bird widely reported from early Coromandel sites and appears to and/or rat bone in D layer 2, very small occasional fragments have been a particular target at Port Jackson, where remains of bone in D layer 3x, and many fragments of bird and/or of at least 50 individuals were identified (Davidson 1979: rat bone in D layer 4. He could make no observations on 188, 191). No such targeting is apparent at Sarah’s Gully. the other samples. His notes are sufficient to suggest that the Nothing much can be said about butchering or other Sarah’s Gully dogs were being fed at least partly on scraps possible practices such as the use of feathers (see B.F. Leach from their masters’ table, so to speak. Mammal remains from the various excavations 1979: 120), as the sample is too small and the retention were studied in detail by Smith (1981). His results are methods unclear. Of the little blue penguins (the most summarised in Table 15. His layer groupings may not numerous birds), the most common identified bones, adequately reflect minimum numbers, as it can be argued complete or partial, at the Settlement were the tibiotarsus that there was really only one occupation layer at each (16), femur (13), coracoid (12) and humerus (8). Other of Areas C and D, where minimum numbers may be bones represented were the ulna (1 adult, 1 juvenile), the inflated; while at Area AB it might be better to group ilium (2), cervical and thoracic vertebrae (2 each), sternum together layer 1 and 2 as level I, layer 3 as Level II, and the (1 juvenile, not from the same context as the other juvenile), other layers as Levels III, IV and V. 140 Tuhinga, Number 29 (2018)

Table 15 Minimum numbers of dogs and sea mammals at Sarah’s Gully (from Smith 1981: 114). Level groupings in Area AB are as follows: level I = layers 1–3; level II = layers 4–5; Level III = layers 6–7; Level IV = layers 8, 9 and natural.

New Zealand Southern New Zealand Assemblage Dog fur seal Cetacean elephant seal sea lion Total Area AB Level I 2 – 1 – – 3 Level II 1 – – – – 1 Level III 3 – 1 – – 4 Level IV 2 – 1 – – 3 Subtotal 8 – 3 – – 11 Area C Layer 3 1 – 1 – – 2 Layer 4 2 1 1 – – 4 Layer 5 2 1 1 – – 4 Subtotal 5 2 3 – – 10 Area D Surface 1 1 1 – – 3 Layer 1 1 – – – – 1 Layer 2 1 – – – – 1 Layer 3 1 – 1 – – 2 Layer 4 2 1 – 1 – 4 Subtotal 6 2 2 1 – 11 Sarah's Midden Layers 2 and 3 1 – – – – 1 Layers 3 and 4 1 1 – – 1 3 Subtotal 2 1 – – 1 4 Pä Layer 2 1 – – – – 1 Layer 3 1 – 1 – – 2 Layer 4 1 – – – – 1 Subtotal 3 – 1 – – 4 Total 24 5 9 1 1 40 Archaeological Excavations at Sarah’s Gully 141

Smith comments on the relatively high number of Tuatara immature dogs, as reflected in epiphyseal fusion and tooth wear, a fact also commented on by Herd. The absence of The easily recognisable mandible of the tuatara (Sphenodon punctatus) was identified only in Area D. One was from layer the New Zealand fur seal (Arctocephalus forsteri) from Area 3A, close to one of the burials, and two were from layer AB is notable, since remains of the species were present 4. These bones no longer appear to be in the collection. in all the other beach middens and almost as numerous as dogs at the Opito Beach Midden. It should be remembered, however, that the area excavated in the lower layers at Area Fish AB was very small compared with the excavation at the Fish from Sarah’s Midden and the 1960 excavation at the Opito Beach Midden. Fur seals were also absent from the Pä (squares 13a–18a and adjacent baulks) were recently rather different sites of Skipper’s Ridge and the Pä. There identified using the methods originally outlined by Leach are no dog or seal remains from the inland part of Area D (1986) and widely followed since. No fish remains appear to at Sarah’s Gully, although Golson’s notes mention scattered have survived from the 1959 excavations at the Pä. In both mammal bones in square 7A on the lowest inland terrace. sites, snapper (Pagrus auratus) was by far the most common Smith subsequently examined further mammal fragments species, followed by leatherjackets (Parika scaber) and, in the from the Pä, identifying additional dog remains from each case of the Pä, labrids (Pseudolabrus spp.) (Table 16). layer, in six different contexts, including those of a very small The limited extent and simple stratigraphy of Sarah’s individual (foetal or a tiny pup), probably from Layer 3, and Midden means that the fish can be readily summarised one additional sea mammal fragment. Six other fragments according to layer (Table 17). Most of the fish are from from three additional contexts were unidentifiable. layers 3 and 5; the few snapper from layers 2 and 4 are The fur seals from the Settlement comprised two probably displaced from those layers. Closer inspection of juveniles and two seals of indeterminate age. Juvenile fur the distribution of fish remains in the Pä is more difficult, as seals were also present at the Opito Beach Midden, Arthur many of them come from the fills of pits. Black’s Midden and Sarah’s Midden; only one pup was identified – at the Opito Beach Midden (Smith 1981: 117). Table 16 Minimum numbers of fish from Sarah’s Midden Smith’s calculation of meat weights from the settlement and the Sarah’s Gully Pä. was confined to his Level IV in Area AB (the earliest cultural Species Sarah’s Midden Pä layer), and layer 3 in Area C. These contexts he found No. % No. % unusual in his study, since here the dog remains comprised Snapper (Pagrus auratus) 41 62.12 32 43.84 only the parts that might be expected to be removed Leatherjacket (Parika scaber) 11 15 and discarded during the butchering process, rather than 16.67 20.55 the prime eating cuts. This may partly explain Herd’s Labrid (Pseudolabrus sp.) 3 4.55 10 13.70 observation about the discard of mandibles (Appendix 6). Kahawai ( trutta) 1 1.52 4 5.48 Bones of the Polynesian rat or kiore (Rattus exulans) Sharks, skates and rays 1 1.52 3 4.11 were collected from all three areas of the Settlement and (Elasmobranchii) catalogued, but have not been studied. They were from Trevally (Pseudocaranx dentex) 3 4.55 1 1.37 the following contexts (number of bones in parentheses; Gurnard – – 3 4 11 * indicates mandibles): Area AB, layer 1 (1*), layer 3 (1*), (Chelidonichthys kumu) . layer 6 (10), layer 8 (3, including 1*), layer 9 (1), layer 10 Porcupine fish (Diodontidae) 2 3.03 – – (21), top of natural (93, including 3*); Area C, layer 3 (1*), Teleostomi species A – – 2 2 74 layer 4 (1*), layer 5 (1); Area D, surface (17, including 1*), (unidentified) . layer 2 (fragments, including 1*), layer 4 (1*). The large (Latridopsis ciliaris) 1 1.52 1 1.37 number from the base of the Area AB deposits is yet another Barracouta (Thyrsites atun) 1 1.52 1 1.37 indication that this was not a pristine area at the time of first Common conger eel 1 – – occupation. (Conger verrauxi) 1.52 Nine rat bones were collected from layer 5 in square Blue cod (Parapercis colias) – – 1 1.37 101 of Sarah’s Midden. Five unsided humeri are listed, Common trumpeter – – giving an MNI between 3 and 5. No rat bones appear to ( lineata) 1 1.52 have survived from the Pä if, indeed, any were collected. Total 66 100 73 100 142 Tuhinga, Number 29 (2018)

Table 17 Minimum numbers of fish by layer at Sarah’s Midden.

Species Layer 2 Layer 3 Layer 4 Layer 5 Total Snapper (Pagrus auratus) 1 10 6 24 41 Leatherjacket (Parika scaber) – 1 – 10 11 Labrid (Pseudolabrus sp.) – – – 3 3 Kahawai () – – – 1 1 Sharks, skates and rays (Elasmobranchii) – 1 – – 1 Trevally (Pseudocaranx dentex) – 1 – 2 3 Porcupine fish (Diodontidae) – 2 – – 2 Blue moki (Latridopsis ciliaris) – – – 1 1 Barracouta (Thyrsites atun) – – – 1 1 Common conger eel (Conger verrauxi) – – – 1 1 Common trumpeter () – – – 1 1 Total 1 15 6 44 66

All fish remains collected from the Settlement A, B, C and D – presumably the left and right dentary and were individually catalogued and variously described premaxilla; ‘upper and lower pharyngeal plates’ of labrids; as jaws, vertebrae, fins, special bones, fragmentary and and leatherjacket spines. Minimum numbers of snapper undistinguished bones, and scales. Golson had consulted are noted as the largest number of A, B, C or D in any the then Dominion Museum curator of fishes, Jock context, and the MNI is therefore probably inflated. The Moreland, who advised him that fish bones were impossible largest number of one bone was 36, from layer 7 of square to identify. However, distinctive bones of three taxa are H10–11 in Area AB. The only other possible identification identified in the catalogue, perhaps on the advice of Ron was a register entry of ‘1 very large spine ? groper’ from Scarlett: four types of ‘jaw bone’ of snapper, described as layer 2 of Area AB.

Table 18 Minimum numbers of fish from Area AB at Sarah’s Gully Settlement (*includes one from the fill of the drain).

Snapper Leatherjacket Labrid Layer (Pagrus auratus) (Parika scaber) (Pseudolabrus sp.) Other Total 1 12 1 3 1 17 2 17 4 1 1 23 3 *8 – – – 8 4 3 2 – – 5 5 2 – – – 2 6 19 2 – 1 22 7 39 4 4 1 48 8 10 2 1 2 15 9 19 4 3 3 29 10 1 – – – 1 10a 3 – – – 3 Total 143 19 12 9 173 Archaeological Excavations at Sarah’s Gully 143

Table 19 Minimum numbers of fish from Area C of the Settlement.

Species Layer 2 Layer 3 Layer 4 Layer 5 Total

Snapper (Pagrus auratus) 1 10 6 23 40 Leatherjacket (Parika scaber) – 1 – 10 11 Labrid (Pseudolabrus sp.) – – – 2 2 Kahawai (Arripis trutta) – – – 1 1 Sharks, skates and rays (Elasmobranchii) – 1 – – 1 Trevally (Pseudocaranx dentex) – 1 – 2 3 Porcupine fish (Diodontidae) – 3 – – 3 Blue moki (Latridopsis ciliaris) – – – 1 1 Barracouta (Thyrsites atun) – – – 1 1 Common conger eel (Conger verrauxi) – 1 – – 1 Common trumpeter (Latris lineata) – – – 1 1 Total 1 17 6 41 65

The results for Areas AB and C are presented in Tables C, and the surface collection in Area D; one porcupine 18 and 19. The entries for Area D are less precise, but fish to layer 4 in Area C; and two tarakihi( Nemadactylus include estimates of five snapper and five ‘other jaws’ macropterus) and an eagle ray (Myliobatis tenuicaudatus) to from layer 2, three snapper from layers 3 and 3A, and 63 layer 5 in Area C. This brings the Settlement even more snapper and one labrid from layer 4. The surface collection into line with the other two sites. included an unspecified number of snapper, ‘1 ?snapper’, Despite the apparently strikingly different evidence of five leatherjackets, one stingray and one unidentified ‘jaw’. fishing technology between the Settlement and the other As snapper, followed by leatherjackets and, in the Pä, two sites – a profusion of one-piece hooks in the former labrids were the most numerous fish identified in Sarah’s and very little fishing gear in the other two – the resulting Midden and the Pä, there are reasonable grounds for catches appear to be similar. There is slight evidence for comparison. It is notable, however, that fewer ‘jaw bones’ trolling only at Area AB and Sarah’s Midden. Apart from other than snapper were listed in the catalogue, suggesting the stone sinkers from the Pä and Sarah’s Midden, which a narrower range of species than in the other two sites; the could well be net sinkers, there is no evidence of nets, but ‘special bones’ may have been mostly other cranial bones of most or all of the fish from these three excavation sites (and the most numerous species, snapper. some of the birds, as noted above) may have been caught in A small sample of distinctive fish bones (other than nets. Fish may also have been caught in traps. snapper) was included in the collection returned to Snapper were the principal species caught by pre- Auckland. These were recently identified. They include a European Mäori in most sites studied on the Coromandel leatherjacket spine and some labrid bones already identified Peninsula and other sites in the northern North Island in the catalogue, and add some labrid premaxillae, which (Leach 2006: 78–79). The emphasis on leatherjackets do not affect the minimum numbers already given. The and sometimes also labrids is typical of some other early range of species is expanded by the addition of one kahawai Coromandel sites, such as Hot Water Beach and Hahei (Arripis trutta) to each of layer 2 in Area AB, layer 3 in Area (ibid.: 337, 338). 144 Tuhinga, Number 29 (2018)

Shellfish D. There were very few shells from Area C and almost none from the upper layers there. The single bag from Area Photographs in the Anthropology Photographic Archive D inland (not included in Table 20) contained minimum at the University of Auckland show that there were no numbers of 10 tuatua (Paphies subtriangulata), two dark ‘concentrated’ shell midden deposits at the three Sarah’s rock shells (Haustrum haustorium) and one common cat’s Gully sites. Shells were collected from virtually all contexts eye. In Table 21, the column headed ‘Features’ groups in the coastal excavations at the Settlement and Sarah’s together shells from the fills of a number of pits and two Midden and from the area of the Pä excavated in 1960, but large postholes. retention strategy was not recorded for any location. The shells are grouped according to soft-shore and rocky- The shells from Sarah’s Gully Settlement were identified shore habitats.1 The rocky-shore species would probably and counted in Auckland before Golson left for Australia all have been available in the immediate vicinity, at either in 1961. A list compiled in Canberra gives the catalogue end of the beach. The soft-shore species are a more diverse numbers of all shells according to species. These analyses group, coming both from exposed beaches, and more tidal yielded a long list of species but relatively few examples of and muddy habitats. The former, particularly tuatua, could each in any context, giving the impression that excavators probably have been gathered in the immediate vicinity. were asked to collect examples of as many different kinds Sarah’s Gully has an open, somewhat exposed sand beach. of shell as possible. These shells had been returned to the The longer beaches at Otama and Opito would also have Auckland Museum and were re-examined against the been a good source of this species. Several species, particularly detailed lists of identifications from the previous analyses. A large dog cockle (Tucetona laticostata) and morning star few shells are now missing and two previously unanalysed (Tawera spissa), which are normally found beyond the low- bags were found. It was apparent that fragments had often tide limit, may have been dead shells collected for purposes been counted during the initial analyses, inflating minimum other than food: Tucetona as a raw material for ornaments numbers, particularly of minor species. Shells from the Pä (Furey 1986) or scrapers (Nichol 1988: 392, fig. 9.13E); and and Sarah’s Midden had not previously been analysed. the others perhaps also for decorative purposes. The single For the present analysis, minimum numbers of bivalves lucine (Divaricella cumingi) specimen would also have been are based largely on sided valves with intact hinges, a dead shell collected on the beach. Although dead shells although in a few cases fragments that cannot be part of of queen scallop (Pecten novaezelandiae) may have been any other counted shell have been included. The greater collected, live shellfish of this species are known to wash number of left or right valves was taken in each context; up at Sarah’s Gully from large beds offshore during rough this may have inflated minimum numbers. In the case of seas (L. Furey, pers. comm., n.d.). Some shells, including gastropods, fragments have been included only if they are chitons, ribbed slipper shells (Maoricrypta costata) (which not part of other counted shells. The larger number of normally live on mussels) and the smallest gastropods, shells or opercula in each context has been taken in the case probably arrived at the site attached to other, larger shells. of Cook’s turban (Cookia sulcata) and common cat’s eye The vermetid snail (Novastoa lamellose) would also have (Lunella smaragda). The more brittle shells, such as Cellana, been an incidental arrival. A single specimen of silver päua Haliotis and Perna, were very fragmented and numbers (Haliotis australis) from layer 3 of Sarah’s Midden is heavily are almost certainly underestimated. During the analysis, encrusted with small barnacles (not listed in Table 22). attempts were made to distinguish radiate limpet (Cellana Despite the small numbers and the doubts about retention radians) from Cellana denticulata, and päua (Haliotis iris) strategy, there are some suggestions of chronological from silver päua (H. australis), but the fragmentary nature change. In the Settlement, soft-shore species, particularly of the material meant that this was unreliable and possibly the tuatua, are relatively more numerous in upper than misleading. Cellana denticulata is definitely present in the lower layers. The principal rocky-shore species are far lowest layers of Area AB, but less clearly present in other more common in lower layers. On the present evidence, it deposits. A similar problem arose with Diloma aethiops and cannot be argued that there is a particular decline in Cellana D. zelandica, which have also been grouped together as denticulata at Sarah’s Gully, rather than a relative decline similar shells from similar habitat. The majority appear to in rocky-shore species generally. Rowland (1976: 9) came be D. zelandica. to a similar conclusion about the Sarah’s Gully limpets. Tables 20–22 summarise the identifications. In Table 20, The general impression provided by the shells from the upper and lower layers are listed separately in Areas AB and lower layers in the Settlement is of rather indiscriminate Archaeological Excavations at Sarah’s Gully 145

Table 20 Minimum numbers of shells from the three areas of the Settlement (* indicates uncertain context). Area AB, Area AB, Area D, Area C, Area D, Species layers upper Subtotal layers Subtotal Total 1–3 4–10 all layers layer 4 Terrestrial Rhytida greenwoodi – – – – 1 – 1 1 Subtotal – – – – – – 1 1 Soft shore Southern olive (Amalda sp.) – 4 4 1 – 2 3 7 Cockle (Austrovenus stutchburyi) 3 – 3 – – – – 3 Lucine (Divaricella cumingi) – – – – – – – 1* Dosinia (Dosinia sp.) 1 – 1 – – 1 1 2 Dog cockle (Glycymeris modesta) 1 1 2 1 – 1 2 4 Screw shell (Maoricolpus roseus) – – – – – 2 2 2 Pipi (Paphies australis) 8 – 8 12 1 1 14 22 Paphies sp. 1 – 1 1 1 – 2 3 Tuatua (Paphies subtriangulata) 39 1 40 12 3 3 18 58 Queen scallop (Pecten novaezelandiae) 5 1 6 4 – – 4 10 foot (Pelicaria papulosa) 1 – 1 – – – – 1 Purple cockle (Purpurocardia purpurata) – 1 1 1 1 – 2 3 Morning star (Tawera spissa) 3 – 3 5 – 3 8 11 Large dog cockle (Tucetona laticostata) 4 5 9 4 6 10 19 Oblong venus (Venerupis largillierti) – – – – – 1 1 1 Xymene plebejus – – – – 2 – 2 2 Turret shell (Zeacolpus fulminatus) – 2 2 1 – – 1 3 Ambiguous trophon (Zeatrophon ambiguous) – 1 1 – 1 1 2 3 Subtotal 66 16 82 42 16 15 72 155 Rocky shore Limpet (Cellana sp.) 5 4 9 68 8 7 83 92 Speckled whelk () 3 – 3 – – – – 3 Cook’s turban (Cookia sulcata) – 2 2 8 5 1 14 16 White whelk (Dicothais orbita) 6 1 7 4 1 – 5 12 Top shell (Diloma sp.) 3 – 3 26 – 2 28 31 Chiton – – – 3 – – 3 3 Päua (Haliotis sp.) 5 2 7 2 – – 2 9 Dark rock shell (Haustrum haustorium) 8 5 13 27 13 1 41 54 White whelk (Haustrum lacunosum) 1 1 2 – 2 – 2 4 Ribbed slipper shell (Maoricrypta costata) 2 – 2 1 – – 1 3 Tiger shell (Maurea punctulata) – 1 1 – – – – 1 Black nerita (Nerita melanotragus) 8 1 9 141 3 1 145 154 Common mussel (Perna canaliculus) 3 – 3 24 – 2 26 29 Rock oyster (Saccostrea glomerata) – – – 8 – 1 9 9 Shield shell (Scutus breviculus) – – – – – 1 1 1 Slit limpet (Tugali elegans) 1 – 1 – – – – 1 Common cat’s eye (Lunella smaragda) 9 5 14 88 2 14 104 118 Wheel shell (Zethalia zelandica) – – – 2 1 – 3 3 Subtotal 54 22 76 402 35 30 467 543 Total 120 38 158 444 51 45 540 699 146 Tuhinga, Number 29 (2018)

Table 21 Minimum numbers of shells from the Pä.

Species Layer 2 Layer 3 Layer 5 Features Total

Soft shore Cockle (Austrovenus stutchburyi) – – – 1 1 Dog cockle (Glycymeris modesta) 1 – – 1 Screw shell (Maoricolpus roseus) – 2 – – 2 Pipi (Paphies australis) 2 1 1 3 7 Tuatua (Paphies subtriangulatum) 18 60 1 82 161 Paphies sp. 1 – – – 1 Queen scallop (Pecten novaezelandiae) 4 1 – 8 13 Small ostrich foot (Pelicaria vermis) – – – 2 2 Morning star (Tawera spissa) – – – 1 1 Large dog cockle (Tucetona laticostata) 3 1 – 1 5 Oblong venus (Venerupis largillieriti) – – 1 – 1 Subtotal 28 66 2 99 195 Rocky shore Limpet (Cellana sp.) – 14 – 54 68 Trumpet shell (Charonia sp.) 1 – – – 1 Cook’s turban (Cookia sulcata) 2 4 1 8 15 Whelk (Dicathais/Lepsithais) 7 6 – 13 26 Top shell (Diloma sp.) 9 17 – 3 29 Chitons 1 – – – 1 Päua (Haliotis sp.) 1 3 2 5 11 Dark rock shell (Haustrum haustorium) 16 11 – 10 37 Tiger shell (Maurea sp). 1 – – – 1 Black nerita (Nerita melanotragus) – 1 – 5 6 Vermetid snail (Novostoa lamellosa) – – 3 – 3 Common mussel (Perna canaliculus) – 1 – 5 6 Rock oyster (Saccostrea glomerata) – – – 2 2 Common cat’s eye (Lunella smaragdus) 22 5 – 42 69 Subtotal 60 62 6 147 275 Total 88 128 8 246 470 Archaeological Excavations at Sarah’s Gully 147

Table 22 Minimum numbers of shells from the Pä.

Species Layer 2 Layer 3 Layer 4 Layer 5 Total

Soft shore Arabic volute (Alcithoe arabica) 1 – – – 1 Cockle (Austrovenus stutchburyi) – – – 1 1 Dog cockle (Glycymeris modesta) – – – 1 1 Pipi (Paphies australis) 1 4 4 2 11 Tuatua (Paphies subtriangulatum) – 10 5 1 16 Paphies sp. – – – 1 1 Queen scallop (Pecten novaezelandiae) – 2 1 1 4 Ostrich foot (Struthiolaria papulosa) – – 1 – 1 Morning star (Tawera spissa) – 1 1 5 7 Large dog cockle (Tucetona laticostata) 1 1 2 1 5 Ambiguous trophon (Zeatrophon ambiguous) – – 1 – 1 Subtotal 3 18 15 13 49

Rocky shore Limpet (Cellana sp.) – – 1 – 1 Cook’s turban (Cookia sulcata) – – 2 1 3 Top shell (Diloma sp.) – – 2 – 2 Päua (Haliotis sp.) – 1 1 – 2 Dark rock shell (Haustrum haustorium) – – – 1 1 Ribbed slipper shell (Maoricrypta costata) – – – 2 2 Rock shell (Neothais/Lepsithais spp.) – 1 1 – 2 Black nerita (Nerita melanotragus) – – 1 4 5 Vermetid snail (Novostoa lamellosa) – – – 2 2 Common mussel (Perna canaliculus) – – – 3 3 Rock oyster (Saccostrea glomerata) – – – 1 1 Shield shell (Scutus breviculus) – – 1 – 1 Common cat’s eye (Lunella smaragda) 2 4 2 4 12 Subtotal 2 6 11 18 37 Total 5 24 26 31 86 148 Tuhinga, Number 29 (2018)

harvesting from the nearby rocks, with some larger species, were primarily artefact collectors, with little interest in such as Haustrum, and a greater number of the medium to context. The only comprehensive archaeological picture of small Turbo and small Nerita and Diloma species, and other early New Zealand prehistory was a South Island one, based small gastropods. largely on Roger Duff’s work (1942, 1947, 1949, 1950, The assemblage from Sarah’s Midden, like those from 1956) at the Wairau Bar moa-hunter site in Marlborough. Areas C and D in the Settlement, is too small to be seen as Prehistoric sequences had been based on attempts to derive anything other than a selection of species present. Although a single coherent account for the whole of New Zealand the two Paphies species together dominate, there is also a from a mélange of Mäori tribal and regional oral traditions good range of rocky-shore species, particularly in layers (e.g. Hiroa 1950: chapters 1–4). 4 and 5. Similarly, in the larger assemblage from the Pä, As noted above, Golson set out to investigate a North tuatua is by far the most numerous single species, but is Island moa-hunter site or sites for comparison with those outnumbered in all layers by the total number of shells from in the South Island. His work at Sarah’s Gully and Opito the rocky shore. showed that some moa had been hunted in the area, but In sum, despite the lack of information on sampling not on any scale. Much more controversial was his claim and retention, it appears that shell gathering at the three that the pits in the inland part of Area D at Sarah’s Gully, excavation sites at Sarah’s Gully was probably primarily which were thought to be storage pits for kümara, had focused on the local rocky shore at all periods, although been contemporary with the moa-hunter occupation with an increasing emphasis on tuatua at later times. There on the beach front. This ran counter to the belief at the does not appear to have been any concentrated gathering time that moa hunters had been just that – hunters – and and processing of soft-shore species. that horticulture had been introduced by a second wave of migrants associated with the ‘Great Fleet’ of Mäori Summary tradition. Roger Green (1963b: 12), in his review of the prehistoric sequence in New Zealand, had argued that the The picture that emerges from the faunal analysis is one of earliest settlers did not have horticulture. Golson (1965) small groups of people periodically harvesting a wide range responded in a well-argued and important paper on the of resources and probably having relatively little impact on role of theory in New Zealand archaeology. The issue their environment. Much of their protein came from the sea: caused considerable discussion for quite a few years, until fish, predominantly snapper, but a range of other species; the existence of early horticulture was clearly confirmed by and shellfish from both rocky- and soft-shore habitats. They radiocarbon dates of garden systems, notably in Palliser Bay also enjoyed meat from their dogs, occasional sea mammals (H.M. Leach 1979), rather than relying on interpretations and a range of birds, mostly seabirds but also some land of inadequately dated pits, assumed to be for the storage of birds. They had no need to concentrate on particular kümara. resources or process food for later consumption elsewhere While the details of Golson’s excavations at Sarah’s – they were free to enjoy the bountiful resources of what Gully, Opito and elsewhere remained largely unpublished, must have seemed a very pleasant place to be, particularly his legacy was his theoretical contribution to the rejection of during the summer. oral traditions as a framework for organising archaeological evidence (Golson 1960), the development of sequences, and his emphasis on the need to understand regional variation Discussion and conclusion (Golson 1959a). Golson’s research at Sarah’s Gully and Opito took place at Sadly, most of the heated debates of the early 1960s a time when important changes were happening in New boiled down to terminological squabbles over terms like Zealand archaeology (Golson 1959a; Golson & Gathercole Moa-hunter, Archaic and East-Polynesian (Duff 1963). 1962). New ideas and techniques had been introduced and But the ground was laid for the exploration of regional applied, initially by Golson in the northern North Island sequences and the consideration of economic, social and from his base at Auckland University College, followed soon environmental factors, in addition to the obsession with afterwards by Peter Gathercole at the University of Otago. moa bones and stone adzes. Previously, archaeological work in New Zealand had been Recent archaeological research at Wairau Bar has shone the province of a tiny number of museum directors and the spotlight again on that site as perhaps the central place curators and a larger number of amateurs, most of whom of the first settlers of New Zealand (Blundell 2014). It has Archaeological Excavations at Sarah’s Gully 149

been suggested that small exploring and colonising parties (Ahuahu), at Whangamata and near Whitianga show that ranged far and wide from Wairau Bar, with some people there is still much that can be learned in this region, even periodically returning there and even being buried there. though important sites are coming under increasing threat The early deposits at the Sarah’s Gully sites (including from coastal erosion and ongoing subdivision of coastal the Pä) can certainly fit such a picture, for none seems to land. Sadly, attempts to secure permanent protection of have been a place of extensive or sustained occupation. the important archaeological landscape of the Kuaotunu Rather, they were the scene of occasional visits by people Peninsula, especially the region of Tahanga, Opito and enjoying the rich natural food resources of the area, Sarah’s Gully, have been unsuccessful. growing crops for a season or two, and taking advantage The significance of the Tahanga basalt resource, of the most important natural resource of all: the Tahanga unknown at the time of the excavations described here, basalt. Sarah’s Gully was just a small part of a scene that has been increasingly recognised following an initial report included not only other sites at Opito and elsewhere on the by Shaw (1963), a brief unpublished excavation by Green Kuaotunu Peninsula, but sites up and down the coasts of and me in 1965 at the request of R.G.W. Jolly, and more the Coromandel Peninsula and nearby islands. detailed accounts by Moore (1975, 1976), Best (1975, Significantly, the later deposits, too, are very similar. 1977) and Turner (Turner 1992; Turner & Bonica 1994). Apart from one stylistically late fishhook point from both The importance of Tahanga in attracting people to the area the Pä and the Settlement, and a lack of extinct bird bones cannot be overemphasised. from upper deposits, there is little in either artefacts or fauna It seems increasingly unlikely that the pits in Area D to differentiate early from late. Later visitors had probably were, in fact, contemporary with the early occupations of not travelled to the Coromandel from Wairau Bar, but they the beachfront areas nearby. But this does not negate the were nonetheless continuing an established tradition of likelihood that the earliest visitors to the area were growing mobility, albeit within a much smaller territory. crops and storing them in pits. The early radiocarbon date Two recent papers argue for rapid of moa from Skipper’s Ridge on charcoal from short-lived species by a small founding human population. Holdaway et al. (Davidson 1974) remains the best evidence for early pit (2014) suggest that moa became extinct in the early to mid- storage on the Kuaotunu Peninsula, and suggests that the fifteenth century AD. Although their paper is primarily early date for a level 3 pit on the Pä may also be correct. concerned with the South Island, they suggested that Evidence of early pit storage has also been reported from there is no convincing evidence of human occupation in Hahei, further south on the Coromandel east coast (Harsant the North Island before the now precisely dated Kaharoa 1983, 1984). Periodic use of storage pits throughout eruption (AD 1314 ±12; Hogg et al. 2003).2 They further the sequence at Sarah’s Gully and Opito is therefore argue that the human population of New Zealand would documented at Sarah’s Pä, Skipper’s Ridge I, T10/777, not have exceeded 2000 people by the time moa became Area D inland, and Skipper’s Ridge II. extinct. Perry et al. (2014) distinguish between a time when Material culture shows little change through time. Use hunting stopped and total extinction (150 and 200 years of the Tahanga basalt seems to have continued throughout after human arrival, respectively). They assume a slightly the sequence and there are very few items, even in the most earlier human arrival, in the thirteenth century, on the basis recent deposits, that would be considered typically ‘Classic of dated rat-gnawed seeds. Both papers support the view of Mäori’. A change in fishing technology from bone to shell a small, highly mobile human population during the first one-piece hooks is well documented at Cross Creek but one to two centuries of occupation of the country. The not reflected in other sites. This, in turn, suggests that other relative scarcity of moa remains at Sarah’s Gully is consistent changes may have been felt that are not yet documented. with these views. Apart from the absence from later layers of the few There has been considerably more research at Sarah’s extinct birds found earlier, there is little indication of change Gully and Opito since 1960, including renewed excavation in the exploitation of a fairly wide range of birds and fish. at Skipper’s Ridge (Bellwood 1969), Sewell’s study of the Only a slight shift in preference from rocky- to soft-shore Cross Creek Midden (Sewell 1984, 1988; Furey et al. 2008), shellfish can be seen, with rocky-shore species continuing and the recent excavation of a large ridgetop settlement at in importance throughout the sequence, and little or no Opito (Bickler et al. 2014). These investigations add support indication of intensive processing of soft-shore species. to the picture that was already emerging from Golson’s In other words, the Sarah’s Gully and Opito sites appear excavations. Recent investigations on Great Mercury Island to reflect a regional sequence that is hardly a sequence at all. 150 Tuhinga, Number 29 (2018)

The appearance of a ditch and bank fortification, albeit a during the last few years, generously providing me with his rather symbolic one, certainly echoes a major and important photographs of Sarah’s Gully, and replying promptly and change felt in most of the country during the middle of the patiently to a barrage of questions. prehistoric sequence. The only other slight indication of Foss Leach negotiated with Jack Golson for the return important change is the unconfirmed possibility that two of the Sarah’s Gully and other archaeological collections extended burials at Sarah’s Gully and Opito are earlier than to New Zealand in 1976, and has helped me with many the crouched and bundled burials of the coastal part of Area D. aspects of the project, including illustrations, radiocarbon This is an important outcome for research that took dates, stone tool analysis and faunal analysis. place at a time of intense interest in ‘the change’ from Ron Scarlett carried out the initial identification of faunal Moa-hunter to Classic Mäori. It reinforces the importance remains; Trevor Worthy has subsequently helped greatly of exploring regional sequences. On the Kuaotunu with bird bones; Ian Smith studied the mammal remains; Peninsula (as at Motutapu Island), early adze technology A.W.B. Powell, many years ago, and Bruce Marshall, much was maintained in the vicinity of an important natural more recently, assisted with shell identifications; and Foss resource for which it was best suited, while new styles and Leach guided me through the fish bone identifications. techniques developed elsewhere. It could be argued that the Rod Wallace has once again brought his expertise to bear Coromandel was a backwater. But the counterargument on charcoal samples, and made an additional contribution would be that people in the Coromandel had no need on silicified wood. P.R. Moore has provided insights into to change. They were in occupation of a region that was unworked stone. ideally suited to their way of life. Artefact illustrations were mostly drawn by Jane Perry The other important aspects of the research at at the then National Museum in Wellington in the late Sarah’s Gully, I believe, are the scope and nature of 1980s‒early 1990s; recently, Joan Lawrence has provided the investigations, which were unprecedented in New final versions of Lawrie Birks’ plans and section of the Pä. Zealand at that time; the training they provided for those Alan Hogg and Fiona Petchey of Waikato University who took part; and the survival of the material for future and Christine Prior, formerly of GNS, assisted with study. At least 1340 m2 were officially hand-excavated at radiocarbon dating. Louise Furey kindly shared information Sarah’s Gully and a further 205 m2 at the two sites at Opito, about the redating of some of the Sarah’s Gully samples, and according to good practice at the time, and fully recorded has been supportive of the project in other ways as well. I in notes, plans, sections and photographs. That there are a am grateful for her thorough and helpful comments on the few missing details after almost 60 years is hardly surprising; manuscript. Two referees provided helpful and perceptive that so much material and documentation has survived comments. trans-Tasman trips, transport within New Zealand and Last but not least, on behalf of all those who have worked less than ideal storage conditions is to be celebrated. at Sarah’s Gully and Opito from 1956 to the present day, Retention strategy certainly fell far below what would be I would like to acknowledge the generosity and assistance required today, but that is hardly surprising. There are still of the late Skipper Chapman and his wife Joyce, and his relatively few sites that have a comparable amount of well- daughter and son-in-law, Sue and Murray Edens. documented faunal and artefactual material available for ongoing study. Notes 1 The speckled whelk (Cominella adspersa) is common on Acknowledgements mudflats as well as rocky shores, but the soft-shore shells at This paper has been long in the making and a great many Sarah’s Gully are almost all from open beaches, rather than people have helped with it over the years. First, I must estuaries. acknowledge the help and inspiration I received from Jack 2 They appear to be unaware of the evidence from Cross Golson while I was an undergraduate student at Auckland Creek (Furey et al. 2008). University. Lawrie and Helen Birks became kind friends and mentors. Roger Green provided constant encouragement and the benefit of his own knowledge of Sarah’s Gully; Wal Ambrose has been a wonderful source of assistance Archaeological Excavations at Sarah’s Gully 151

References Davidson, J.M. and Leach, B.F. (2017). Archaeological excavations at Pig Bay (N38/21, R10/22), Motutapu Beever, R.E., Parris, B.S. and Beever, J.E. (1969). Studies on Island, Auckland, New Zealand, in 1958 and 1959. the vegetation of Cuvier Island. 1. The plant communities Records of the Auckland Institute and Museum 52: 9–39. and a species list. Tane 15: 53–68. Davidson, J.M. and Green, R.C. (1975). A locality map for Bellwood, P. (1969). Excavation at Skipper’s Ridge, Opito Skipper’s Ridge (N40/7), Opito. Records of the Auckland Bay, Coromandel Peninsula, North Island of New Institute and Museum 12: 43–46. Zealand. Archaeology and Physical Anthropology in Duff, R.S. (1942). Moa-hunters of the Wairau. Records of Oceania 4(3): 198–221. the Canterbury Museum 5: 1–43. Best, S. (1977). The Maori adze: an explanation for change. Duff, R.S. (1947). The evolution of native culture in New Journal of the Polynesian Society 86(3): 307–337. Zealand: moa-hunters, , Maoris. Mankind 3(10, Best, S.J. and Merchant, R.J. (1976). Siliceous sinter and 11): 281–92, 313–22. the early Maori. New Zealand Archaeological Association Duff, R.S. (1949). and man (part I). Antiquity 23: Newsletter 19: 106–109. 172–179. Birks, L. (1960). Pä at Sarah’s Gully. New Zealand Duff, R.S. (1950). Moas and man (part II). Antiquity 24: Archaeological Association Newsletter 3(2): 16–20. 72–83. Birks, L. and Birks, H. (1970). Radiocarbon dates for a pä Duff, R.S. (1956). The Moa-hunter period of Maori culture. site at Sarah’s Gully. New Zealand Archaeological Wellington: Government Printer. xix + 400 pp. Association Newsletter 13(2): 63. Duff, R.S. (1963). New Zealand archaeology. Antiquity 37: Birks, L. and Birks, H. (1973). Additional dates for Sarah’s 65–68 [Notes and News]. Gully pä site. New Zealand Archaeological Association Edson, S. and Brown, D. (1977). Salvage excavation of Newsletter 16(2): 73. an Archaic burial context, N44/97, Hahei. Records of the Blundell, S. (2014). Where it all began. New Zealand Auckland Institute and Museum 14: 25–36. Listener, 1 January. Falla, R.A., Sibson, R.B. and Turbott, E.G. (1970). A Boileau, J. (1980). The artefact assemblage from the Opito field guide to the birds of New Zealand. 2nd edn. London: Beach Midden, N40/3, Coromandel Peninsula. Records Collins. 256 pp. of the Auckland Institute and Museum 17: 65–95. Fisher, V.F. (1936). Ethnologist’s report. In: Annual report Calder, A. and Calder, J. (1977). A head-ache or a pain of the Auckland Institute and Museum for 1935–1936. in the neck? A small-scale excavation of a midden/burial Auckland: Auckland Institute and Museum. Pp. 18–19 at Opito Bay, Coromandel. New Zealand Archaeological and plate. Association Newsletter 20(1): 39–40. Furey, L. (1986). Maori pendants made from dog cockle Davidson J.M. (1974). A radiocarbon date from Skippers shells. New Zealand Archaeological Association Newsletter Ridge (N40/7), Opito, Coromandel Peninsula. New 29(1): 20–28. Zealand Archaeological Association Newsletter 17: 50–52. Furey, L., Petchey, F., Sewell, B. and Green, R. (2008). Davidson, J.M. (1975). The excavation of Skipper’s Ridge New observations on the stratigraphy and radiocarbon (N40/7), Coromandel Peninsula, in 1959 and 1960. dates at the Cross Creek Site, Opito, Coromandel. Records of the Auckland Institute and Museum 12: 1–42. Archaeology in New Zealand 51(1): 46–64. Davidson, J.M. (1979). Archaic middens of the Coromandel Golson, J. (1955). New Zealand Archaeological Association. Region: a review. In: Anderson, A. (ed.). Birds of a Journal of the Polynesian Society 64: 349–351. feather. Osteological and archaeological papers from the South Golson, J. (1957). Field archaeology in New Zealand. Pacific in honour of R.J. Scarlett. New Zealand Journal of the Polynesian Society 66(1): 64–109. Archaeological Association Monograph 11 and BAR Golson, J. (1959a). Culture change in prehistoric New International Series 62. Oxford: British Archaeological Zealand. In: Freeman, J.D. and Geddes, W.R. (eds). Reports. Pp. 183–202. Anthropology in the South Seas. : Avery. Davidson, J.M. (1984). The prehistory of New Zealand. Pp. 29–74. Auckland: Longman Paul. 270 pp. Golson, J. (1959b). Excavations on the Coromandel Davidson, J.M. (1986). The poor man’s rei puta: a shell Peninsula. New Zealand Archaeological Association pendant from Sarah’s Gully. New Zealand Archaeological Newsletter 2(2): 13–18. Association Newsletter 29(4): 224–227. Golson, J. (1960). Archaeology, tradition and myth in New 152 Tuhinga, Number 29 (2018)

Zealand prehistory. Journal of the Polynesian Society Leach, B.F. (1979). Maximizing minimum numbers: avian 69(4): 380–402. remains from the Washpool Midden Site. In: Anderson, Golson, J. (1965). Some considerations of the role of theory A. (ed.). Birds of a feather. Osteological and archaeological in New Zealand archaeology. New Zealand Archaeological papers from the South Pacific in honour of R.J. Scarlett. Association Newsletter 8(2): 79–92. New Zealand Archaeological Association Monograph Golson, J. and Gathercole, P. (1962). The last decade in 11 and BAR International Series 62. Oxford: British New Zealand archaeology. Antiquity 36: 168–174, 271– Archaeological Reports. Pp. 103–121. 278. Leach, B.F. (1986). A method for the analysis of Pacific Green, R.C. (1963a). Summaries of sites at Opito, island fishbone assemblages and an associated data base Sarah’s Gully, and Great Mercury Island. New Zealand management system. Journal of Archaeological Science Archaeological Association Newsletter 6(1): 57–69. 13(2): 147–159. Green, R.C. (1963b). A review of the prehistoric sequence Leach, B.F. (2006). Fishing in pre-European New of the Auckland province. New Zealand Archaeological Zealand. New Zealand Archaeological Association Association Monograph 2. Auckland: New Zealand Special Publication and Archaeofauna 15. Wellington: Archaeological Association. New Zealand Archaeological Association. iv + 359 pp. Groube, L.M. (1993). Dig up those moa bones, dig. Golson Leach, H.M. (1979). Evidence of prehistoric gardens in in New Zealand, 1954–1961. In: Spriggs, M., Yen, eastern Palliser Bay. In: Leach, B.F. and Leach, H.M. D.E., Ambrose, W., Jones, R., Thorne A. and Andrews, (eds). Prehistoric man in Palliser Bay. Bulletin of the A. (eds). A community of culture. The people and National Museum of New Zealand 21. Wellington: prehistory of the Pacific. Occasional Papers in Prehistory, National Museum of New Zealand. Pp. 137–161. 21. Canberra: Department of Prehistory, Research Leahy, A. (1991). Excavations at Taylor’s Hill, R11/96, School of Pacific Studies, Australian National University. Auckland. Records of the Auckland Institute and Museum Pp. 6–17. 28: 33–68. Hamilton, A., Campbell, K., Rowland, J. and Browne, Moore, P.R. (1975). Preliminary investigation of the P. (2016). The Kohuamuri siliceous sinter as a vector Tahanga basalt, Coromandel Peninsula. New Zealand for epithermal mineralisation, Coromandel Volcanic Archaeological Association Newsletter 18(1): 32–36. Zone, New Zealand Mineralium Deposita. International Moore, P.R. (1976). The Tahanga basalt: an important Journal for Geology, Mineralogy and Geochemistry of stone resource in North Island prehistory. Records of the Mineral Deposits 126: 1–16. Auckland Institute and Museum 13: 77–93. Harsant, W. (1983). Radiocarbon dates from N44/97, Moore, P.R. and Wallace, R. (2000). Petrified wood from Hahei, Coromandel Peninsula. New Zealand the Miocene volcanic sequence of Coromandel Archaeological Association Newsletter 26(1): 59–61. Peninsula, northern New Zealand. Journal of the Royal Harsant, W. (1984). Archaic storage pits at N44/97, Hahei, Society of New Zealand 30(2): 115–130. Coromandel Peninsula, New Zealand. New Zealand New Zealand Archaeological Association (2009). ArchSite: Journal of Archaeology 6: 23–35. Archaeological Site Recording Scheme [website]. Hiroa, T.R. (Buck, P.) (1950). The coming of the Maori. Retrieved 7 June 2016, from www.archsite.org.nz. 2nd edn. Wellington: Maori Purposes Fund Board. Perry, G.W.L, Wheeler, A.B., Wood, J.R. and Wilmshurst, 551pp. + xxiv pls. J.M. (2014). A high-precision chronology for the rapid Hogg, A.C., Higham, T.G.F., Lowe, D.J., Palmer, J., extinction of New Zealand moa (Aves, Dinornithiformes). Reimer, P. and Newnham, R.M. (2003). A wiggle Quaternary Science Reviews 105: 126–135. match date for Polynesian settlement of New Zealand. Rowland, M.J. (1976). Cellana denticulata in middens on Antiquity 77: 116–125. the Coromandel Coast, NZ – possibilities for a temporal Holdaway, R.N., Allentoft, M.E., Jacomb, C., Oskam, C.L., horizon. Journal of the Royal Society of New Zealand Beavan, N.R. and Bunce, M. (2014). An extremely low- 6(1): 1–15. density human population exterminated New Zealand Rowland, M.J. 1977. Seasonality and the interpretation of moa. Nature Communications 5. doi: 10.1038/ncomms6436. the Tairua Site, Coromandel Peninsula, NZ. Archaeology Houghton, P. (1977). Human skeletal material from and Physical Anthropology in Oceania XII(2): 135–150. excavations in eastern Coromandel. Records of the Scarlett, R.J. (ed.) (1957). Members’ doings. New Zealand Auckland Institute and Museum 14: 45–56. Archaeological Association Newsletter 1(1): 3. Archaeological Excavations at Sarah’s Gully 153

Scarlett, R.J. (ed.) (1958). Editorial. New Zealand Unpublished sources Archaeological Association Newsletter 1(3): 1. Scarlett, R.J. (1979). Avifauna and man. In: Anderson, Best, S. (1975). Adzes, rocks and men. M.A. research essay, A. (ed.). Birds of a feather. Osteological and archaeological University of Auckland, Auckland. papers from the South Pacific in honour of R.J. Scarlett. Bickler, S., Baquié, B., Brown, A., Clough, R., Dawson, New Zealand Archaeological Association Monograph L., Hand, A., Judge, C., et al. (2014). Opito Bay 11 and BAR International Series 62. Oxford: British T10/777, Coromandel Peninsula: archaeological Archaeological Reports. Pp. 75–90. excavation. Unpublished report in fulfilment of NZHPT Sewell, B. (1986). Radiocarbon dates from the Cross Creek Authority No. 2008/85, Clough and Associates, site (N40/260), Sarah’s Gully, Coromandel Peninsula. Auckland. New Zealand Archaeological Association Newsletter 29(4): Newton, Z.J. (2000). Waitaia silica sinter: occurrence and 228–229. aspects of genetic history. B.Sc. Hons thesis, University Sewell, B. (1988). The fishhook assemblage from the of Auckland, Auckland. Cross Creek site (N40/260; T10/399), Sarah’s Gully, Nichol, R. (1988). Tipping the feather against a scale: Coromandel Peninsula, New Zealand. New Zealand archaeozoology from the tail of the fish. Ph.D. thesis, Journal of Archaeology 10: 5–17. University of Auckland, Auckland. Sewell, B. (1990). Opito and Otama sites revisited. Scarlett, R.J. (n.d.). Handwritten list of bone identifications Archaeology in New Zealand 33(4): 189–202. (principally birds) from Sarah’s Gully, sent to Jack Shaw, E. (1963). Maori quarry, Tahanga hill. New Zealand Golson; copy sent to Roger Green. Pp. 9–14 of a larger Archaeological Association Newsletter 6(1): 34–36. manuscript. Smith, I.W.G. (1981). Prehistoric mammalian fauna from Seelenfreund-Hirsch, A. (1985). The exploitation of Mayor the Coromandel Peninsula. Records of the Auckland Island obsidian in prehistoric New Zealand. Ph.D. thesis, Institute and Museum 18: 107–125. Anthropology Department, University of Otago, Turner, M.T. and Bonica, D. (1994). Following the flake Dunedin. trail: adze production on the Coromandel east coast, Sewell, B. (1984). The Cross Creek site (N40/260), New Zealand. New Zealand Journal of Archaeology 16: Coromandel Peninsula. M.A. thesis, Department of 5–32. Anthropology, University of Auckland, Auckland. University of Waikato (n.d.). New Zealand radiocarbon Turner, M.T. (1992). Make or break: adze manufacture at database [online database]. Retrieved 10 July 2015, from the Tahanga quarry. M.A. thesis, University of Auckland, www.waikato.ac.nz/cgi-bin-nzcd. Auckland. Worthy, T.H. (1999). What was on the menu? Avian extinction in New Zealand. New Zealand Journal of Archaeology 19(1997): 125–160. Worthy, T.H. and Holdaway, R.N. (2002). The lost world of the moa. Prehistoric life of New Zealand. Christchurch: Canterbury University Press. 718 pp. 154 Tuhinga, Number 29 (2018)

Appendix 1: Details of pits in Area D inland, Sarah’s Gully Settlement

Pit 1: oblong Pit 6: rectangular 4 ft x 2 ft 3in (122 x 68 cm). Maximum depth, 18 in (45 6 ft 6 in x 3 ft–3 ft 6 in (198 x 91–106 cm). Maximum cm). Internal ledge at downslope end only. Depth of ledge, depth in natural, 8 in (20 cm); minimum depth, 3 in (17 4 in (10 cm). Shallow posthole-like depression in one cm). Slight lip around edge. No features. Bottom fill of corner of ledge. Partly filled, and sealed, by layer C. natural silting, covered by layers A and B. Partly sealed by layer C. Pit 2: square (Golson 1959a: plate 1C) 3 ft x 3 ft (91 x 91 cm). Maximum depth into natural on Pit 7: rectangular upslope side, 2 ft 4 in (71 cm). Maximum depth into natural 10 ft 6 in x 4 ft 6 in (320 x 137 cm). Maximum depth, 8 in downslope side, 18 in (45 cm). Maximum depth below (20 cm); minimum depth, 6 in (15 cm). External shelf near ledge, 18 in (45 cm). Ledge significantly wider and deeper southwest corner. Depths of central postholes from north at downslope end. Thin black layer on floor. Feature in to south, 6 in (15 cm), 4 in (10 cm), 10 in (25 cm). Depth corner of ledge labelled ‘minute post base’. of external postholes from east to west, 6 in (15 cm) and 3 in (7 cm). Thin dark layer with earth and sand on pit floor. Pit 3: rectangular Partly sealed by layer C. (Golson 1959a: plate 1B rear) 6 ft 2 in x 2 ft 10 in (187 x 86 cm). Maximum depth in Pit A: rectangular natural, 15 in (38 cm); depth minimum, 9 in (22 cm). Two 4 ft 3 in x 1 ft 11 in (129 x 55 cm). Maximum depth in rounded subsurface ‘bins’ dug into and below pit wall in natural, 7 in (18 cm); minimum depth, 3 in (7 cm). Depth one corner. Maximum depth of bin floor, 20 in (50 cm) of postholes, 13 in (33 cm) at centre of upslope wall, 7 in below top of natural. Posthole depths, east to west, 9 in (18 cm) in pit floor. Drain from Pit 4 appears to run into (22 cm), 1 ft 5 in (43 cm), 12 in (30 cm). Thin layer of corner of pit. Partly sealed by layer C. clean, water-washed sand on floor. Drain 3 in (7 cm) deep in floor, 13 in (33 cm) below natural at exit, c. 11 in (28 cm) Pit B: rectangular at 12f–g boundary. 5 ft 6 in x 3 ft (167 x 91 cm). Maximum depth in natural, 3 in (7 cm). One central posthole, 9 in (23 cm) deep. Pit 4: rectangular (Golson 1959a: plate 1B foreground) Pit C: rectangular (?) 6 ft x 3 ft 3 in (182 x 99 cm). Maximum depth in natural, 12 >4 ft x 2 ft 4 in (122 x 71 cm). Maximum depth, 3 in (7 in (30 cm); minimum depth, 8 in (20 cm). Depths of three cm). No recorded features. main postholes, 17 in (43 cm). Depth of small posthole in north wall, 2 ft 5 in (73 cm). Thin layer of clean, water- Feature 37: rounded square washed sand on floor. Drain 3 in (7 cm) deep in pit floor, 2 ft 9 in (84 cm) diameter. Maximum depth into natural, 6 11 in (27 cm) in north face of square 12f, 7 in (17 cm) at in (15 cm). No features. Sealed by layer X. west edge of square 12g.

Pit 5: rectangular 5 ft 5 in x 3 ft 6 in (165 x 106 cm). Maximum depth in natural, 18 in (45 cm); minimum depth, 12 in (30 cm). No postholes. Drain present and shallow in about half pit length, 18in (45 cm) deep in natural at exit, dropping to 6 in (15 cm) 12 ft (3.65 m) away. There was a long, shallow depression in the surface of the pit fill, sealed by layer B, with several fill layers, including one with lumps of weathered rhyolite. Archaeological Excavations at Sarah’s Gully 155

Appendix 2: Details of pits at Sarah’s Pä Rectangular pits Square 14, baulk 14–14A Pit A North–south. 221 x 99 x 28 cm. Slightly narrower Orientation of the main group of pits is described as north– at north end. Shallow postholes, one at south end, one at south if the long axis followed the central spine of the ridge, middle of east side. One posthole central, 35 cm deep; two transverse if at right angles, and diagonal if neither of the other postholes in floor off centre line, 33 cm and 20 cm other two. Depth is the depth of the pit floor below the deep. surface of the natural into which it was dug, rather than the level from which it was actually dug. Three shallow Square 14a, baulks 14–14b, 14b–15b structures in squares 16B to 18B were described by Birks in Pit A North–south. 198 x 137 x 28 cm. One posthole (not his field notes as pits, but had little or no wall on the downhill central) in each of east, south and west walls, two in midline sides. They followed the side of the ridge (although not the of floor. contour, rising from northeast to southwest) and were thus Square 15A, baulk 15A–15 diagonal to the main group. Pit B North–south. 102 x 81 x 30 cm. Uneven floor. Three probable central postholes. Squares 11A–12A, including baulk 11A–12A Unnumbered Transverse. Rounded rectangular pit, 213 x 130 cm. Depth not recorded. Three central postholes down long axis, 28 cm, 15 cm and 30 cm deep. One andesite slab resting in hole against south wall, one in pit fill (Figs A2.1 and A2.2).

Squares 12a–13a, including baulk 12a–13a Pit A North–south. Rounded rectangular pit, 130 x 91 x 30 cm. Probably but not definitely more recent than Pit X.

Pit X North–south. 270 x 114 cm. Depth not recorded. Shallow, with floor rounded up at sides. Three central postholes in middle part of pit floor; three postholes outside along southern edge and two on northern edge.

Squares 13a–14a, including baulk 13a–14a Pit B (square 13a)/Pit C (square 14a) Diagonal. 175 x 69 x 20 cm. One posthole 40 cm deep, central, but close to southern end; one shallow posthole, probably intrusive from layer 2, in southeast corner.

Fig. A2.1. The transverse rectangular pit in squares 11A to 12 A. The hole on the right contained a large block of andesite.

Fig. A2.2. The stone in its original position in the pit. A second similar stone was in the fill on the opposite side of the pit. 156 Tuhinga, Number 29 (2018)

Square 16, baulks 16–16A, 15–16 Pit A 236 x 84 cm (north end), widening to 99 cm (south Pit A North–south. 198 x 137 x 28 cm. At least four central end); 18 cm deep on uphill side, no wall remaining on postholes (Fig. A2.3). downhill side. Continuous with Pit B but set down about 10–15 cm at the junction of the two. One central posthole in floor near each end wall; southern one not vertical so that post would incline towards centre of pit. Pit B 190 x 99 cm (north end), tapering to 69 cm (south end). Four postholes on midline of floor, four along downhill long edge. A semi-circular shelf, c. 45 x 45 cm, set into the uphill wall 10 cm above pit floor, with two postholes aligned at right angles to the pit wall. Uphill wall 18 cm, no wall remaining on downhill side. The southeast corner of Pit B was about 40 cm from the northwest corner of the extension to Pit C. Pit C 305 x 140 cm, with a partial lengthwise extension, 61 x 69 cm, at the northeast corner; 23 cm deep on the uphill side and as little as 10 cm deep on the downhill side. Five shallow postholes (6–19 cm deep), irregularly aligned on the floor; three postholes (9–13 cm deep) evenly spaced on the downhill rim. Fig. A2.3. The rectangular pit (A) in square 16. Bin pits Square 16, baulk 16–16a These are rectangular pits, but considerably smaller than the Pit B Transverse. 160 x 76 x 38 cm. One posthole in the rest of the rectangular pits. Orientation is as described for midline of the floor towards the west end, one against the the rectangular pits. centre of the east wall. A 20 cm-deep, 30 cm-diameter hole in the floor on the north side, undercutting the pit wall. Square 16 and baulks 15–15a–16–16a Unnumbered Transverse. Sharply rectangular with straight Squares 16B–18B sides. 93 x 68 x 40 cm. One central posthole, 20 cm deep Three shallow aligned pits (Fig. A2.4): (Fig. A2.5).

Fig. A2.4. The three adjoining rectangular “pits” in Fig. A2.5. The bin pit in square 16 and adjacent baulks. squares 16B to 18B. A: foreground, B: centre, D: rear. Archaeological Excavations at Sarah’s Gully 157

Square 15a ‘Grave pits’ Unnumbered Diagonal. 46 x 30 x 18cm Square 11 Unnumbered Diagonal. Oval plan, slightly undercut at each Bell-shaped pits end. 127 x 33 (maximum) x 53 cm. No postholes (Fig. A2.6).

Square 14a, baulk 14–14a Pit A1 Irregular rounded square mouth, c. 120 cm wide; 79 cm deep. Slightly undercut all round.

Square 15 and baulk 15–16 Pit A Mouth diameter 46 cm, depth 107 cm, floor diameter 68 cm.

Pit B Mouth diameter 53 cm, maximum depth 46 cm, floor slanting upwards to undercut south side.

Pit C Irregular mouth, 76 x 53 cm, depth not recorded. Floor sloping at about 20 degrees south to north, undercut on north side only.

Pit D Irregular mouth, 84 x 56 cm, depth not recorded, curved floor, slightly undercut on southwest side.

Pit E Oval mouth, 64 x 38 cm, floor 64 x 61 cm, heavily undercut on northwest side only.

Square 15 and baulk 15–26 Pit A Oval shape, slightly undercut. 102 x 81 x 30 cm. Fig. A2.6. The “grave” pit in square 11. Baulk 14–15 Unnumbered 96 x 58 x 30 cm. Square 14, baulk 14–14a Unnumbered Diagonal. Oval plan. 127 x 51 (maximum) x 40 cm. No postholes. Appendix 3: Burials in the coastal part of Area D, Sarah’s Gully Settlement R.J. Scarlett Formerly of Canterbury Museum, Christchurch

Burial 1, 1a. Platform D lower first premolar, first and second right molars, and first and Bones of this mixed burial had weathered out. Later, a few third left molars were still in the mandible when recovered. bones were recovered from the ashy midden layer, and it There was little wear on the cusps, except for the first was learned that Arthur Black had already found some bones (‘fern-root’) molars, which were considerably worn. Other and reburied them. Other miscellaneous bones were also long bones, being considerably smaller, much smoother and found adjacent to Burial 2. All these were probably from the less callused at the proximal and distal ends, were almost one burial. They consisted of a number of vertebrae, ribs, certainly of an adult female, designated Burial 1a. To this phalanges, long bones, a mandible and cranial fragments. individual can be added a frontal fragment, with part nasals The mandible and some of the long bones (broken) were and part palate, from which the teeth had been lost after those of a young adult male (designated Burial 1). All teeth death. This bone showed signs of being considerably older had been present at death, and the incisors, canines, right than those of Burial 1. 158 Tuhinga, Number 29 (2018)

Burial 2. Platform D lower Burial 3 (originally Burial X). (Figs A3.1 and A3.2) Platform D upper This was a trussed burial. The grave had been dug down A few phalanges, ribs and other fragments from this burial through ashy midden sand into clean yellow sand and had been found weathered down the slope. The grave had surrounded by a ring of stones. The cranium had sunk been dug through the ashy midden into clean sand, and forward onto the breast. The limbs were drawn up sideways. surrounded by a ring of stones. No long bones, cranium or The cranium was of an adult male. The maxilla lacked an mandible were present. The body was an adult, probably incisor (probably lost after death). The mandible showed male. The excavated remains consisted of the pelvis (in strong evidence of caries. The first and second molars and pieces), vertebrae, ribs, scapulae and a few digits. These three premolars were absent. Most of the missing teeth bones were not in a position of articulation. were probably lost long before death as the bone had healed over the gaps. The surviving cusps were well worn. The Burial 4. Platform D lower, mandible had a pronounced chin. The sutures were well facing sea (Fig. A3.3) defined and partly fused. The age was estimated at 25–40 The cranium had been uncovered by Arthur Black. It or possibly more. had been disturbed and the mandible was missing. The remainder of the crouched, trussed burial was undisturbed. It was an adult male aged 30 or more. The cranial sutures were well defined, but partly fused. Both third molars had been lost before death; the second molars on each side were well worn. The first molars and second premolars were very large. The two front incisors had been lost after death. All teeth were larger than normal, and the cusps were well worn. The maxilla was fairly prognathous. This skeleton lay in clean yellow sand.

Fig. A3.1. Burial 2, Area D.

Fig. A3.2. Burial 2, Area D.

Fig. A3.3. Burials 4 to 6 and 6a, Area D. Archaeological Excavations at Sarah’s Gully 159

Burial 5. Platform D lower Burial 7 This was in darker sand, adjacent to Burial 4. The cranium Adult male. All bones except the mandible were present. and mandible were absent. The body was that of an adult The cranial sutures were well fused but visible. The base male, crouched and trussed and placed in a partly supine had been removed, but the frontals and right zygomatic position. Lengths of long bones were: femur, 18 in [46 cm]; arch were present. Age estimated as 30 or more. Long bone tibia, 15 in [38 cm]; humerus 12½ in [32 cm]. lengths: humerus, 15 in [38 cm]; ulna, 11 in [28 cm]; femur, 17¼ in [44 cm]; tibia, 20 in [51 cm]. Burial 6. Platform D lower This adult male was below Burial 5, in the darker sand, in a Burial 8 different orientation to Burial 5. The body was only partly Adult female represented by the cranium only, cracked excavated, but did not appear to be trussed. The cranium into fragments after burial (? by pressure). The walls were and mandible were not present. thin with advanced fusion of sutures. The base had been removed. No other bones. Burial 6a. Platform D lower This was a large young adult male, lying in clean sand, Burial 9 adjacent to Burial 4. The body was only partly excavated. Cranium of a child, cracked in the same way as Burial 8, The cranium and mandible were not present. with the base removed. Neck vertebrae and a few ribs were present. Possibly 8–10 years old. Burials 7, 8 and 9 (originally designated Burial Y). Burial 10. Platform D middle Platform D lower (Fig. A3.4) This was in a hole about 30 cm square and was similar to This group burial, probably a family group, was initially Burial 3, except that no ring of stones was present. This exposed by Arthur Black. The three individuals were tightly burial consisted of a collection of disarticulated bones, packed together in a pit about 2 ft [60 cm] in diameter, dug including vertebrae, pelvis (no sacrum) and ribs, but no through the black sand into the yellow sand, and about 2 ft cranium or long bones. The individual was an adult, sex [60 cm] or 3 ft [90 cm] distant from Burial 2. not determined.

Fig. A3.4. Burials 7 to 9, Area D. 160 Tuhinga, Number 29 (2018)

Appendix 4: Silicified wood from Sarah’s Gully Rod Wallace Department of Anthropology, University of Auckland, PO Box 92019, Auckland ([email protected])

Eight bags containing pieces of rock recovered from References Area AB of the Settlement and Sarah’s Pä that had Furey, L., Phillipps, R., Emmitt, J., Jorgensen, A., been tentatively identified as silicified wood were Ladefoged, S. and Holdaway, S. (2017). Brief interim received for examination. All are the same material; report for excavations on Ahuahu Great Mercury Island it was determined to be silicified wood with a well- 2014–2017. Archaeology in New Zealand 60(3): 45–63. preserved cell structure that is readily observable under Moore, P.R. and Wallace, R. (2000). Petrified wood from the microscope (Fig. A4.1, lower). This material the Miocene volcanic sequence of Coromandel Peninsula, northern New Zealand. Journal of the Royal was originally wood from a conifer species. The fine Society of New Zealand 30(2): 115–130. anatomical details are insufficiently preserved to allow identification to taxon. Although six of the samples were simply unmodified rock fragments, two had been worked to some degree by grinding and polishing one or more faces (Fig. A4.1, upper). Silicified wood is quite abundant on the Coromandel Peninsula and surrounding islands (Moore & Wallace 2000). It was formed following the burial of forests by volcanic tephras of largely Miocene age. On Great Mercury Island (Ahuahu), a little to the northeast of Sarah’s Gully, it is ubiquitous in sediments and is sufficiently common to have been used by Mäori as oven stones. Silicified wood is frequently found in a form that preserves much of the grain of the original wood, so that it readily fractures into long, parallel-sided rods, as is the case here. In one site excavated in 2014 on Great Mercury Island (Ahuahu) (Furey et al. 2017), a fishing lure was found made from material identical to the pieces illustrated here. In the same site, five other rod-shaped pieces, some with minor polished facets like the pieces here, were also recovered; these were clearly manufacturing blanks. A polished rod-shaped pendant of identical material was also surface-collected on the island. I am not suggesting that the Sarah’s Gully material necessarily originated from Great Mercury Island (Ahuahu), although this is a possibility, but clearly it was being brought to the Sarah’s Gully sites as a useful raw material. Fig. A4.1 Silicified wood pieces and image of one surface showing the cross section of the wood cell structure. Archaeological Excavations at Sarah’s Gully 161

Appendix 5: Sarah’s Gully charcoal and vegetation Rod Wallace Department of Anthropology, University of Auckland, PO Box 92019, Auckland ([email protected])

The use of results from identification of charcoal in from distant areas will have been collected from beaches reconstructing vegetation surrounding sites during their adjacent to sites and used, as will wood from old logs and occupation is not always straightforward. Much of the stumps left behind from forests that were cleared well before charcoal will be residue from firewood collected daily by the sites were occupied. the inhabitants from dead wood gathered in the immediate Charcoal samples collected from fills of pits, postholes vicinity of settlements. As such, it potentially represents a and other features representing built structures typically repeated sampling of the local vegetation present at the time contain a somewhat different range of species from those the sites were occupied. Other firewood sources, however, from cooking features that contain only firewood remains. are likely to have been exploited. Driftwood that originated The former may include remains of building timbers

Table A5.1 All charcoal identifications from Sarah’s Gully.

Settlement Pä Sarah’s Midden Species Plant type Pieces % Pieces % Pieces % Bracken (Pteridium esculentum) Fern 1 0.2% – – – – Shrub spp. 6 4 – Tutu (Coriaria sp.) 31 10 – Hebe (Hebe sp.) 55 39 1 Coprosma (Coprosma sp.) 90 50 19 Fivefinger (Pseudopanax arboreus) 2 6 – Pittosporum (Pittosporum sp.) 4 10 1 Olearia (Olearia sp.) 2 1 1 Lancewood (Pseudopanax crassifolius) – – 10 Akeake (Dodonaea viscosa) 19 – – Mingimingi (Leucopogon fasciculatus) 4 – 5 Whärangi (Melicope ternata) Shrub and 4 4 1 64% 82% 45% Puka (Griselinia lucida) scrub species – 1 – Patëē(Schefflera digitata) – 1 – Hangehange (Geniostoma ligustrifolium) – 1 1 Kawakawa (Piper excelsum) – 1 – Ribbonwood or Plagianthus sp. 2 – 3 Ngaio (Myoporum laetum) 21 31 3 Toro (Myrsine salicina) 10 – – Mäpou (Myrsine australis) 2 2 – Mähoe (Melicytus ramiflorus) 5 5 2 Mänuka (Leptospermum scoparium) 29 10 – Känuka (Leptospermum ericoides) – 71 – Tawa (Beilschmiedia tawa) – 1 – Taraire (Beilschmiedia taraire) Broadleaved 1 – – 29% 15% 53% Püriri (Vitex lucens) trees 15 25 22 Pöhutukawa (Metrosideros excelsa) 115 19 33 Kauri (Agathis australis) 10 9 1 Kahikatea (Dacrycarpus dacrydioides) Conifers 22 7% 1 3% – 2% Mataï (Prumnopitys taxifolia) – – 1 Total 449 302 104 162 Tuhinga, Number 29 (2018)

quite possibly obtained from outside the immediate area. Table A5.2 Charcoal identifications from Furthermore, charcoal from post-occupation landscape T11/914, Whitianga. fires tends to accumulate in the fills of abandoned pits and ditches, and will include species such as bracken (Pteridium esculentum), tutu, hebe and coprosma that colonise the Species Plant type No. of % disturbed environments of abandoned sites. Finally, it is quite Pieces common to find veins or zones of charcoal in natural soil Bracken (Pteridium esculentum) 2 0.2% horizons in the sites, often all from a single species, which Hebe (Hebe sp.) 12 are clearly burnt-out tree roots from fires that occurred long Coprosma (Coprosma sp.) 10 before the sites were occupied. Whärangi (Melicope ternata) 1 In 2003, I identified material in 28 charcoal samples (plus two wood samples from postholes) from the Pä and five Dracophyllum 7 from Sarah’s Midden (Wallace 2004). In 2007, I examined (Dracophyllum sp.) 72 bags from the Settlement, 41 of which were sampled Ramarama (Lophomyrtus bullata) 31 (Wallace 2007). The combined results are shown in Table Akeake (Dodonaea viscosa) Shrub and 24 A5.1. 18% Pseudopanax (Pseudopanax sp ) scrub species 6 In general, the charcoal results are dominated by shrub . and scrub species, accompanied by püriri and pöhutukawa as Pittosporum (Pittosporum sp.) 8 the only common broadleaved trees. The latter trees typically Olearia (Olearia sp.) 20 survive forest clearance and remain on cleared and farmed Mäpou (Myrsine australis) 26 landscapes today. There is some minor variation between the sites. The Settlement samples have a little more conifer Ngaio (Myoporum laetum) 7 charcoal (kauri and kahikatea, Dacrycarpus dacrydioides) Mänuka 8 than the other sites, perhaps reflecting the presence of (Leptospermum scoparium) more building timbers. The Pä has a higher proportion of Tawa (Beilschmiedia tawa) 1 känuka and püriri. Sarah’s Midden has a higher percentage Mähoe (Melicytus ramiflorus) 7 of pöhutukawa, probably reflecting its beachfront location. Mänuka and känuka are scrub species that become Tawäpou (Pouteria costata) 13 dominant in vegetation regenerating on landscapes after Täwheowheo (Quintinia serrata) 1 clearance by fire, yet, except at the Pä, even these are Pukatea 1 present in limited amounts. Very frequent firing of the (Laurelia novae-zelandiae) Broadleaved 24% landscape will suppress most woody vegetation, which will Rewarewa (Knightia excelsa) trees 1 be replaced by a cover of bracken, accompanied by a range of small shrubs. The charcoal evidence from all these sites Hïnau/Pökäkä (Elaeocarpus sp.) 4 clearly indicates that forest had been removed long before Maire (Nestegis sp.) 10 they were occupied, and that frequent fires had resulted Püriri (Vitex lucens) 17 in a plant cover of fern and scrub species, with püriri and Pöhutukawa pöhutukawa the only surviving broadleaved trees. (Metrosideros excelsa) 162 The charcoal results from the Sarah’s Gully sites can usefully be compared with those from two other local sites. Kahikatea (Dacrycarpus dacrydioides) 290 The first of these is T11/914, an early site with moa remains at Whitianga (Table A5.2). Here, more than 80% of the Tötara (Podocarpus totara) 13 Conifers 57% charcoal was from 10 species of broadleaved trees and five Rimu (Dacrydium cupressinum) 3 species of conifers. This clearly demonstrates that primary Mataï (Prumnopitys taxifolia) 92 forest was present when the site was occupied. Kauri (Agathis australis) 113 The second site for which data are available is at Opito Mangrove Bay (T10/777), less than 2 km from Sarah’s Gully (Table Estuarine 6 1% A5.3). Three areas of occupation on a ridge behind the beach (Avicennia marina var. resinifera) at the northern end of the bay were investigated (Bickler et Total 896 al. 2014: 150–162, 200–205). At the end of the ridge two Archaeological Excavations at Sarah’s Gully 163

large concentrations of archaeological features were found At Areas 1 and 2, kauri forms 55% of the charcoal from (Areas 1 and 2), with numerous storage pits of varying sizes cooking areas, with shrub species contributing 40% and all and types, postholes suggestive of above-ground structures, other trees contributing only 5%. Although kauri is the most and areas of burning and cooking. A third concentration abundant firewood species present, living coastal broadleaved of small fire scoops and midden (Area 3), to the north, was podocarp forest, such as that indicated for T11/914 at also uncovered. Nine radiocarbon dates indicate occupation Whitianga, contains a diverse collection of large trees, with from the mid-fifteenth to sixteenth centuries in Areas 1 and kauri as only one element. There is no such thing as a living 2, and the mid-nineteenth century in Area 3. The charcoal kauri forest accompanied only by shrubs. The only possible was systematically collected both from structural features situation that could have produced this association of species (mainly fills of pits) and from cooking areas. The results is one where a kauri forest had been destroyed and replaced show charcoal from structural features has a wider range of by shrubs and scrub vegetation, but had left behind logs and species than the firewood remains, indicating that building stumps of the durable resinous wood on the landscape, which timbers were probably present. were collected and used for firewood. Interestingly, the

Table A5.3 Charcoal identifications from site T10/777 at Opito Bay.

Pit fills Firewood Firewood Species Plant type c. AD 1600 c. AD 1600 c. AD 1850 Pieces % Pieces % Pieces % Bracken (Pteridium esculentum) Fern 3 0.3% 1 0.2% – – Monocotyledon 1 1 – Tutu (Coriaria sp.) 11 8 30 Hebe (Hebe sp.) 12 27 76 Coprosma (Coprosma sp.) 18 17 5 Akeake (Dodonaea viscosa) 4 9 2 Patë (Schefflera digitata) Shrub and – – 1 17% 40% 74% Olearia (Olearia sp.) scrub species 26 54 141 Rangiora (Brachyglottis repanda) 15 – – Pseudopanax (Pseudopanax sp.) 1 – 1 Ngaio (Myoporum laetum) 1 – – Mänuka (Leptospermum scoparium) 45 101 7 Mähoe (Melicytus ramiflorus) 33 3 11 Pöhutukawa (Metrosideros excelsa) 61 24 83 Püriri (Vitex lucens) 14 – – Maire (Nestegis sp.) 10 – – Broadleaved Beech ( sp ) 38 18% 3 5% – 28% . forest trees Pukatea (Laurelia novae-zelandiae) 1 – 8 Hïnau (Elaeocarpus sp.) 46 1 – Kohekohe (Dysoxylum spectabile) 3 – – Tötara (Podocarpus totara) 3 – – Mataï (Prumnopitys taxifolia) Conifers 8 65% – 55% 1 1% Kauri (Agathis australis) 628 303 2 Total 982 552 368 164 Tuhinga, Number 29 (2018)

charcoal from the fire scoops and midden in Area 3 dating to around the nineteenth century was mainly from shrubs (74%) and pöhutukawa (23%), indicating that by that time, sub-fossil kauri wood had all been used and the local woody vegetation consisted of shrubs and pöhutukawa. Discussion The charcoal identified from the Sarah’s Gully sites reflects a landscape considerably altered from that found by the earliest arrivals in the area. The original natural vegetation would almost certainly have been closed-canopy forest growing down close to the water’s edge, as is indicated at T11/914 at Whitianga. What is interesting is that the traces of kauri-dominated forest, which is present in the sixteenth century at T10/777 in nearby Opito Bay, are absent from the Sarah’s Gully samples. Most of these were from contexts that are either undated, or known to be relatively recent. The charcoal results strongly suggest that almost all was from later periods in the prehistoric era, when the area supported a vegetation cover of fern, shrubs and scrub, with a fringe of pöhutukawa and some persistent stands of püriri. References Unpublished sources Bickler, S., Baquié, B., Brown, A., Clough, R., Dawson, L., Hand, A., Judge, C., et al. (2014). Opito Bay T10/777, Coromandel Peninsula: archaeological excavation. Unpublished report in fulfilment of NZHPT Authority No. 2008/85, Clough and Associates, Auckland. Wallace, R. (2004). Charcoal identification, Sarah’s Pa and Sarah’s Gully Flat, Opito, Coromandel Peninsula. Unpublished report to Janet Davidson, 8 January. Wallace, R. (2007). Charcoal from Sarah’s Gully Midden and Settlement. Unpublished report to Janet Davidson, 30 November. Tuhinga 29: 165–173 Copyright © Museum of New Zealand Te Papa Tongarewa (2018) 165

Increasing visitor engagement in a contemporary art exhibit through a participatory activity Edith MacDonald

Formerly Head of Visitor Insights, Museum of New Zealand Te Papa Tongarewa, PO Box 467, Wellington 6011, New Zealand. Current position: Manager, Social Science, New Zealand Department of Conservation, PO Box 10420, Wellington 6143, New Zealand ([email protected])

ABSTRACT: Over the last decade, there has been a rise in participatory activities in museums and art galleries. A common form of participatory activity is to ask visitors questions, based on the objects, and their answers remain in the gallery for subsequent visitors to experience. However, there have been few empirical studies to assesses the impact a participatory activity has on visitor experience. This study, carried out at the Museum of New Zealand Te Papa Tongarewa, Wellington, in 2012, compared visitor behaviour (time in exhibition and at objects) and visitor cognition (satisfaction, elaboration, relevancy, perception of the art exhibition) during a participatory and control period. Participants in the activity had significantly greater levels of elaboration and stayed almost three times as long in the exhibition compared to the control period. The activity resonated with a youthful audience, with two-thirds of participants under 30. While there is some concern by museum staff that participatory activities may be frivolous or detract from the art, the findings of this study indicate they are successful at fostering a deeper engagement in visitors.

KEYWORDS: Interpretation, visitor studies, art education.

Introduction are inspired by the objects, remain in the gallery and become an extension of the exhibit for subsequent visitors to Over the last few decades, museums and art galleries experience (Simon 2010). Participatory activities are based have aspired to deepen visitor engagement and have a on the social nature of people (Adams et al. 2004) and build transformative impact on the visitor (Falk & Foutz 2007; upon the societal trend of sharing greater amounts of personal Soren 2009; Henry & McLean 2010; Chang 2012; information and opinions to a wide audience (Madden et McCall & Gray 2014). To facilitate deeper engagement al. 2013). A common participatory activity is to ask visitors with exhibitions, museums and galleries have increasingly questions connected to the object, with visitors posting used interactives to personalise the subject matter, evoke their responses on sticky notes to a public comment board. emotion, stimulate dialogue and promote reflection for Such questions ‘encourage visitors to engage deeply and the visitor (Adams et al. 2003, 2004; Falk et al. 2004; personally with a specific object to motivate interpersonal Patel et al. 2016). An interactive is defined as a range of dialogue among visitors around a particular object or idea’ experiences that fully engage visitors personally, physically (Simon 2010: chapter 4, p. 9). The questions do not have a and emotionally (Adams et al. 2004). Through their use, correct answer but are abstract, and their main purpose is to visitors make their own meaning of the objects or exhibition stimulate interest and engagement in the visitor. by incorporating their pre-existing personal knowledge and Simon (2010) provided galleries and museums beliefs (Sienkiewicz 2015). The use of interactives has been with detailed examples and methodology to develop a linked to significant long-term learning outcomes in visitors participatory exhibition with the goal of engaging visitors (Falk et al. 2004). on a deeper level. As a result, numerous art galleries have Participatory activities are a subset of interactive been experimenting with participatory exhibitions to engage experiences. In participatory activities, visitor contributions with visitors (e.g. Santa Cruz Museum of Art and History, 166 Tuhinga, Number 29 (2018)

Columbus Museum of Art, Oakland Museum of California, are still reluctant to provide visitors with resources to create Carnegie Museum of Art and Milwaukee Art Museum). and post their own content (Patel et al. 2016). While there are several blog posts discussing general positive To remedy the debate about the impact of a participatory outcomes of a participatory activity (e.g. exhibitions activity on the visitor, greater empirical evidence is required. attendance was increased, anecdotal positive feedback This paper fills the gap and evaluates how a participatory art from visitors), there is little empirical evidence to reveal exhibition has a positive behavioural and cognitive impact how a participatory activity impacts visitor engagement on visitors. (Kirchberg & Trondle 2012; Skibins et al. 2012). Only a few theses, grey reports and peer-reviewed publications Method (e.g. Nelson 2011; Greco 2014; Patel et al. 2016) assess the Study site impact of participatory exhibitions, especially those asking Collecting Contemporary was a contemporary art exhibition visitors to post answers to questions, on visitor experiences held at the Museum of New Zealand Te Papa Tongarewa (i.e. cognitive and behavioural changes). If a museum’s (Te Papa) between June 2011 and July 2012 (Museum of intention is to be transformative (Soren 2009), it is necessary New Zealand Te Papa Tongarewa n.d.). The exhibit was to conduct a deeper form of assessment on participatory located on the fifth floor of the museum and extended to experiences, and on interactives in general. approximately 800 m2. Data was collected during the control Participatory activities, and interactives in general, can condition (20 February to 1 March 2012), when there was be a source of dilemma among museum staff. Museums, no participatory activity in the gallery. On 19 March 2012, especially art galleries, want to reach new audiences on a the participatory activity, ‘Stick it to the man’, was installed deeper level and move away from the perception that they (Fig. 1). The activity was branded with a unique colour are uninteresting or that art is difficult to understand (Chang pattern, logo and title to capture the attention of visitors 2012). While interactive and participatory activities may arriving at the entrance to the exhibit. Participatory data was engage visitors and be seen as fun, some museum staff worry collected between 23 April and 29 April 2012. that visitors are just playing and not learning (Adams et al. During the participatory activity, a host (when available) 2003). Museum staff can view the participatory activity as an approached visitors and asked them if they wanted to ‘interloper on sacred ground’ (Adams et al. 2004: 158) that participate in the activity. If they opted in, they were given draws attention away from the object. As a result, museums a clipboard and a short explanation of the activity, and were

Fig. 1 The entrance to the Collecting Contemporary exhibition, with the participatory activity ‘Stick it to the man’ installed (photo: Edith MacDonald). Visitor engagement participatory activity 167

directed to the first object. On the clipboard was a short Relevance was measured because people are influenced explanation of the activity, four colour-coded sticky notes, by messages that have a high bearing to their life (Petty et pictures of the artworks and a pencil. At each participatory al. 2003). Visitors are more likely to elaborate on issues that object, there was a colour-coded question on a wall-mounted are relevant to them and connected to their lives (Petty & sign with the same branded logo seen at the entrance. The Cacioppo 1996). Satisfaction was measured to evaluate the colour coding matched the sticky note on the clipboard, on overall visitor assessment of the exhibit. If an experience which the visitors were instructed to write their response is satisfying, processing of the experience may increase to the question. Upon exiting the exhibit, visitors posted (Wegener et al. 1995). However, if the visitor perceives the each response on the corresponding coloured section of the experience to be dissatisfying, they may disengage with the communal board and returned the clipboard. Nine feeder experience (Petty et al. 2003). responses were on the board. Responses were collected at To measure satisfaction, relevancy and elaboration, the close of each day and dated. I used a commercially available questionnaire based on Based on Simon’s (2010) recommendations for writing communication theory (Ham & Weiler 2006) (questions prompting questions, a team of curators, educators, 4–20 in Appendix 1). This questionnaire has been used interpreters and writers developed questions or prompts for in other informal learning environments (e.g. Weiler & the four objects: Smith 2009; MacDonald et al. 2015). The questionnaire Paratene Matchitt, The family (1966): Tell us about a family took approximately 10 minutes for visitors to complete and memory that springs to mind. consisted of multiple questions measuring three constructs Lisa Walker, Pendant (2009): Describe the strangest gift on a Likert scale of 1 to 7: satisfaction (four questions; α = you’ve ever received. 84), relevancy (four questions; α = 0.67) and elaboration Reuben Paterson, Natural (2005): If you jumped through (five questions;α = 0.83). the centre of this painting, where would you be? In addition to the Likert-scaled questions, respondents Simon Morris, Pause 4.5 and 5.5 (2006): Tell us about the were asked three open-ended questions: what three last rule you broke. words would you use to describe this exhibit? What was your favourite aspect of this contemporary exhibit? If you Questionnaire participated in the sticky note activity, tell us your opinion For the questionnaire, visitors over the age of 18 were of this activity. Participants also provided information on surveyed as they left the exhibit (Appendix 1). Potential their age and group size to determine the impact of those respondents were selected using the next-to-pass method variables, if any, on responses. (the first visitor to pass an imaginary line was approached). In the case of a group, the person selected for the questionnaire Behavioural observations was the visitor closest to the researcher. Refusal rate was tallied. Behavioural observations were conducted over three Based on the recommendation made by Falk & Foutz days during the control condition (2–4 March 2012) and (2007), I selected a survey instrument that could assess participatory condition (4–6 May 2012). Observations were the cognitive impact (if any) of the participatory activity. conducted daily from 11am to 2pm. A research assistant Specifically, I wanted to measure visitor satisfaction, was positioned at the entrance of the exhibit and selected relevance and elaboration during the participatory and a visitor to observe based on the next-to-pass method. control conditions. Elaboration, also known as provocation, The makeup of the group and gender of the visitor were thinking or meaning making, is considered a key outcome recorded. Once a visitor was selected for observation, the of a museum visit (Falk & Foutz 2007). Elaboration leads to research assistant noted their entrance into the gallery, their visitors making a connection between the exhibit experience path on a map of the gallery and their exit time. Visitor and what they already know and feel. An increase in participation in the sticky note activity was noted. Time elaboration leads visitors to process new thoughts about the spent viewing three objects used in the participatory activity experience (Tubb 2003), can lead to a shift in attitude (Petty was recorded, as visitor attention and attraction power are & Cacioppo 1996) or strengthen an existing attitude (Petty important indicators of visitor mindfulness and learning & Wegener 1999), and can guide future behaviour (Bagozzi (Moscardo 1999). The researcher also noted if the visitor et al. 2002). had read the object label. 168 Tuhinga, Number 29 (2018)

Questionnaire analysis Results The dependent variables elaboration, satisfaction and Refusal rate was less than 5% during both the control relevancy were measured using multi-item seven-point and participatory condition. Data was compared between Likert-scaled questions. As justified by Rasmussen (1989), respondents during the control period (control; n = 232), the data were treated as continuous and thus warranted visitors who participated in the ‘Stick it to the man’ activity parametric analysis. The average of each construct was (participants; n = 27, 17% of visitors) and visitors who did calculated and ranged from 1 to 7. Prior to analysis of not participate in this (bystanders; n = 135, 83% of visitors). the dependent variables, I assessed whether there was a The main reason visitors did not participate in ‘Stick it difference in sociodemographic variables (age and size) to the man’ was because they were not invited to do so among the groups. I also calculated Cronbach’s alpha to by a staff member (37%). Only 7% of respondents did not assess the reliability of the constructs. participate because of lack of interest in the activity. Sixty- seven per cent of participants in ‘Stick it to the man’ were under the age of 30, compared to 42% of the control group Table 1 Age breakdown of control, bystanders and and 44% of the bystander group (Table 1). There was no participants in the ‘Stick it to the man’ participatory significant difference p( >0.05, Fisher’s exact test) in place of activity. residence (Wellington region, rest of New Zealand or other country) among the three groups. There was no difference % of % of % of control in group size between bystanders and participants, 2 (4) = Age bystanders participants χ group 6.27, p = 0.108. (n = 232) (n = 135) (n = 27) Participants engaged in ‘Stick it to the man’ because they 18–19 12 7 15 thought it was fun/interesting (38%), wanted to relate to the 20–29 30 37 52 art in a new way (23%) or to help Te Papa (12%). 30–39 16 26 15 There was a significant difference in satisfaction, 40–49 10 4 4 relevancy and elaboration among the three groups, F(2, 391) = 3.46, p<0.005, partial η2 = 0.02 (Fig. 2). 50–59 11 11 15 Participants had significantly greater rates of elaboration and 60–69 15 11 0 relevancy compared to the control group and bystanders

Over 70 5 3 0 (pparticipants<0.05), but did not differ on satisfaction

(pparticipants>0.05).

6 control

♦ bystander 5 ♦ participant

4

3

2

1

0 satisfaction relevancy elaboration

Fig. 2 Mean rate (± standard error) of satisfaction, relevancy and elaboration in the three groups. ♦ = p<0.05. Visitor engagement participatory activity 169

Table 2 Mean rate (± standard error) of interest in art and knowledge of art (scale 1 to 7), and percentage of visitors whose view of contemporary art changed following their visit, across the three groups.

View of contemporary Interest in art Knowledge of art art has changed Control group (n= 232) 5.177 ± 0.10 3.719 ± 0.10 17%

Bystanders (n = 135) 5.008 ± 0.27 3.266 ± 0.30 19%

Participants (n = 27) 5.222 ± 0.13 3.444 ± 0.13 31%

There was no significant difference between the control Discussion group, participants and bystanders in interest of art, F(2, 391) = 0.60, p>0.05. There was a significant difference The results from the study indicate that a participatory in knowledge of art of the three groups, F(2, 391) = exhibition does have a positive cognitive and behavioural 3.37, p<0.05, with the control group having significantly impact on visitors. Visitors who partook in the ‘Stick it greater knowledge in art than the bystander group. As a to the man’ activity showed an increase in elaboration, result of participation, 31% of visitors stated their view of relevance and time spent in the exhibition, and a reduced contemporary art had changed compared to 17% of the negative perception of art. However, only 17% of visitors control group and 19% of bystanders (Table 2). participated in the activity, suggesting that the impact of a The top five words used to describe the exhibit were participatory activity may be limited to a small number of consistent across the three groups, although the percentage visitors. Visitors stated that the main reason they did not varied: ‘interesting’ (control = 20%; bystander = 26%; participate was that they were not asked to do so. Thus, participant = 23%), ‘different’ (control = 8%; bystander to increase the overall impact of a participatory activity, a = 11%; participant = 15%), ‘intriguing’ (control = 6%; staff member or volunteer needs to be at the entrance to bystander = 14%; participant = 11%), ‘modern’ (control = the exhibition to invite visitors to take part in the activity. 4%; bystander = 6%; participant = 8%) and ‘varied’ (control While this may add cost or resources to the overall budget, = 6%; bystander = 9%; participant = 8%). Twenty per cent it is essential if the goal of the participatory activity is to have of the words listed by the control group were negative (e.g. a positive impact on the greatest number of visitors possible ‘boring’, ‘not my scene’, ‘pretentious’), compared to 3% of (Adams et al. 2003). Interestingly, the participatory activity bystanders and 0% of the participants. had a modest impact on bystanders, suggesting that not Respondents who participated in the ‘Stick it to the everyone has to participate actively in order to receive the man’ activity were more inclined to speak with a host in the benefit of a participatory activity. This concept is discussed exhibit (38%), compared to bystanders (11%) or the control below. group (13%). ‘Stick it to the man’ resonated with a youthful audience, Visitors participating in ‘Stick it to the man’ stayed with two-thirds of participants being under the age of significantly longer M( = 18.41 minutes; SD = 3.01) in 30 (more than a 50% increase on numbers of under-30s the exhibit than control visitors (M = 6.06 minutes; SD = compared to the control and bystander groups). The high 0.68) or bystanders (M = 8.17 minutes; SD = 1.42). One proportion of people under 30 may be impacted by the hundred per cent of the participants viewed the three pieces broader cultural trend in which young people are more likely of artwork, compared to 21% of the control group and 19% to share opinions and personal information actively through of the bystanders. Participants viewed the three selected social media (Madden et al. 2013). Thus, participating and objects six times longer (M = 0.60 minutes; SD = 0.28) sharing their opinions in an art exhibit may not be divergent than control (M = 0.07 minutes; SD = 0.04) or bystander to their everyday behaviour. If museums want to engage (M = 0.09 minutes; SD = 0.08) groups. and attract young people, which is a key goal for Te Papa 170 Tuhinga, Number 29 (2018)

(Museum of New Zealand Te Papa Tongarewa 2014), this (Tubb 2003). If museums strive to have a long-term and study indicates that incorporating participatory activities transformational impact on visitors (Soren 2009), then this into the exhibits will be necessary. This is important for study suggests that participatory exhibitions like ‘Stick it to future exhibition development, as many young people feel the man’ will help facilitate this goal. excluded from, and not at home in, art galleries (McCarthy Art galleries, compared to zoos and museums, can be & Mason 2006). In contrast, no visitor over the age of perceived by visitors as places where learning is presented in 60 participated in the activity, despite 20% of the control the least engaging manner and the opportunity to interact group and 14% of the bystanders being in this age group. with the exhibits is limited (Packer & Ballantyne 2002). Despite this lack of participation, the level of satisfaction of This negative perception of art galleries appears to change the exhibit among the over-60s was not impacted by the when visitors participate in an activity. Visitors who partook presence of the participatory activity, suggesting that it did in ‘Stick it to the man’ had reduced negative perceptions not detract or alienate these visitors. of modern art following their experience, as indicated by The main reason visitors participated in ‘Stick it to the negative words listed, from a measure of 20% in the control man’ was to engage in an enjoyable and interesting activity. group to 0% in participants. The reduction of negative Of note, pre-existing interest and knowledge of art neither words to zero suggests that participants could relate to the hindered nor facilitated participation in ‘Stick it to the exhibition in a positive way and perceived its content as man’. This result allays concern that such an activity may be more accessible and relevant to them. offensive to art enthusiasts. On the contrary, it suggests that As noted earlier, those visitors who entered the exhibition participatory exhibitions have a broad appeal across content when ‘Stick it to the man’ was installed but did not personally expertise and interest. There was no difference in reasons participate in the activity (i.e. bystanders) were still impacted for participation between the age groups. These results positively. Bystanders spent almost two minutes more in the support previous research that people want to be stimulated exhibition than the control group and had greater rates of in an entertaining way during their leisure time (Falk et al. elaboration compared to the control group. Furthermore, 1998). The second most cited reason for participation was to their negative perception of an art exhibition was reduced engage in art in a new way. Understanding the motivation to only 3%. It is plausible that, although the bystanders were of visitors who participate is necessary, as it will assist in not active participants in ‘Stick it to the man’, they still designing future programmes to ensure more people are engaged as spectators and received some benefits from the engaged based on their motivational state (Falk et al. 1998; activity. For example, bystanders may still read the questions Packer & Ballantyne 2002). at each artwork or look at the communal response board, Participation in ‘Stick it to the man’ resulted in a even though they do not participate publicly. This form significant increase in time spent in the exhibit and time at of passive participation by observation is common in social each object for participants compared to the control group. media such as Facebook, Twitter and Instagram. At a fundamental level, the more time a visitor spends in the exhibit, the greater the learning (Serrell 1997). Thus, increasing visitor time in an exhibit is desirable. The fact that Conclusion participatory visitors in this study stayed in the exhibition significantly longer than the control and bystander groups The use of sticky notes to ask visitors questions in an suggests that the participatory aspect of the exhibition was exhibition has been utilised by many museums and art successful at changing visitor behaviour. galleries. However, there have been few empirical studies Compared to the control and bystander groups, visitors on the impact of participatory exhibitions on visitors. participating in ‘Stick it to the man’ had an increased rate Results from this study demonstrate the positive impacts the of elaboration, a key cognitive outcome for museums ‘Stick it to the man’ participatory activity had on behaviour (Falk & Foutz 2007). Increased elaboration indicates that and cognitive processing of visitors to the Collecting the participatory activity stimulated in-depth cognitive Contemporary exhibition, and can provide guidance for processing, which may lead to long-term storage of future programming to increase the impact of exhibitions. information and result in a greater impact on the visitor Fundamentally, staff or volunteers should facilitate Visitor engagement participatory activity 171 participatory exhibition activities in order to increase the References number of visitors engaging in it, as only a small number of visitors opt in without being invited. Thus, designing future Adams, M. and Moussouri, T. (2002). Interactive learning in museums of art and design. In: Proceedings of the programmes should incorporate staff/volunteer resources to Interactive Learning in Museums of Art and Design ensure visitors are invited to participate. conference, Victoria and Albert Museum, London, The main reason visitors engaged in the activity was 17–18 May 2002. London: Victoria and Albert Museum. to have fun, a motivation museums need to embrace. Accessible at http://media.vam.ac.uk/media/documents Historically, museums and galleries have grappled with /legacy_documents/file_upload/5748_file.pdf. the balance between education and entertainment, and the Adams, M., Moreno, C., Polk, M. and Buck, L. (2003). The issues of developing experiences that focus on entertainment dilemma of interactive art museum spaces. Art Education at the expense of learning (Adams & Moussouri 2002). 56: 42–52. However, as Falk et al. (1998) note, the dichotomy is Adams, M., Luke, J. and Moussouri, T. (2004). Interactivity: artificially constructed by museums and not of concern to moving beyond terminology. Curator 47: 155–170. visitors. Instead, visitors who list entertainment as a main Bagozzi, R.P., Gurhan-Canli, Z. and Priester, J.R. (2002). motivation for a museum visit learn more and stay at the The social psychology of consumer behaviour. Buckingham: museum longer compared to those who express a high Open University Press. 224 pp. educational motivation. When the team was developing Ballantyne, R. and Uzzell, D. (2011). Looking back and the ‘Stick it to the man’ initiative, many staff members looking forward: the rise of the visitor-centred museum. Curator 54: 85–92. questioned the purpose of the activity, and were concerned Chang, E. (2012). Art Trek: looking at art with young that few people would participate and that it would diminish children. International Journal of Education Through Art the overall visitor experience by ‘dumbing it down’. My 8: 151–167. results confirm that visitors perceive participatory activities Dudzinska-Przemitzki, D. and Grenier, R.S. (2008). as fun, and that the activities increase cognitive processing Nonformal and informal adult learning in museums: and do not compromise the integrity of the exhibition for a literature review. Journal of Museum Education 33(1): visitors who opt not to participate. 9–22. Having empirical evidence to support participatory Falk, J. and Foutz, S. (2007). In principle, in practice: museums initiatives will continue to push museums to be visitor- as learning institutions. Lanham, MD: Altamira Press. centric, allowing them to develop programmes that meet 338 pp. the needs and agendas of visitors and not be moderated by Falk, J., Moussouri, T. and Coulson, D. (1998). The effects potential bias beliefs of staff (Ballantyne & Uzzell 2011). of visitors’ agendas on museum learning. Curator 41: 107–120. Falk, J., Scott, C., Dierking, L., Rennie, L. and Jones, M. Acknowledgements (2004). Interactives and visitor learning. Curator 47: 171–198 I would like to thank the curators, writers, interpreters, Ham, S.H. and Weiler, B. (2006). Development of a research- designer and hosts who supported and contributed to this based tool for evaluating interpretation. Brisbane: CRC project, in particular Sarah Farrar, Frith Williams, Sarah for Sustainable Tourism. 56 pp. Morris, Lucinda Blackley and Steve Brady. Henry, B. and McLean, K. (eds) (2010). How visitors changed our museum: transforming the Gallery of California Art at the Oakland Museum of California. Oakland, CA: OMCA. 144 pp. Kirchberg, V. and Trondle, M. (2012). Experiencing exhibitions: a review of studies on visitor experiences in museums. Curator 55: 435–452. McCall, V. and Gray, C. (2014). Museums and the ‘new museology’: theory, practice, and organisational change. Museum Management and Curatorship 29: 19–35. 172 Tuhinga, Number 29 (2018)

McCarthy, C. and Mason, D. (2006). ‘The feeling of Petty, R.E., Wheeler, S.C. and Tormala, Z.L. (2003). exclusion’: young people’s perceptions of art galleries. Persuasion and attitude change. Pp. 353–382. In: Millon, Museum Management and Curatorship 21: 20–31. T. and Lerner, M.J. (eds). Handbook of psychology. Vol. 5: MacDonald, E., Milfont, T.L. and Gavin, M. (2015). personality and social psychology. New York, NY: Applying the elaboration likelihood model to increase John Wiley & Sons. 688 pp. recall of conservation messages and elaboration by zoo Rasmussen, J. (1989). Analysis of Likert-scale data: a visitors. Journal of Sustainable Tourism 24(6): 1–16. reinterpretation of Gregoire and Driver. Psychological Madden, M., Lenhart, A., Cortesi, S., Gasser, U., Duggan, Bulletin 105: 167–170. M., Smith, A. and Beaton, M. (2013). Teens, social Serrell, B. (1997). Paying attention: the duration and media, and privacy. In: Pew Research Center [website], allocation of visitors’ time in museum exhibitions. online report, 21 May. Retrieved on 8 April 2018 Curator 40: 108–125. from http://assets.pewresearchorg/wp-content/uploads/ Sienkiewicz, N. (2015). Creating meaningful experiences sites/14/2013/05/PIP_TeensSocialMediaandPrivacy_ in art museums. Pp. 231–245. In: Anderson D., Cosson PDF.pdf. A. and McIntosh L. (eds). Research informing the practice Moscardo, G. (1999). Making visitors mindful: principles for of museum educators. Rotterdam: SensePublishers. 256 pp. creating quality sustainable visitor experiences through Simon, N. (2010). The participatory museum. Santa Cruz, effective communication. Champaign, IL: Sagamore CA: Museum 2.0. 388 pp. Accessible online at www. Publishing. 132 pp. participatorymuseum.org. Museum of New Zealand Te Papa Tongarewa (n.d.). Skibins, J.C., Powell, R.B. and Stern, M.J. (2012). Exploring 2011 past exhibitions / Ngä whakaaturanga 2011. In: empirical support for interpretation’s best practices. Museum of New Zealand Te Papa Tongarewa [website]. Journal of Interpretation Research 17: 25–44. Retrieved on 7 April 2018 from www.tepapa.govt.nz/ Soren, B.J. (2009). Museum experiences that change visitors. visit/whats-on/exhibitions/past-exhibitions/2011-past- Museum Management and Curatorship 24: 233–251. exhibitions#Collecting. Tubb, K.N. (2003). An evaluation of the effectiveness of Museum of New Zealand Te Papa Tongarewa (2014). interpretation within Dartmoor National Park in Te Papa statement of intent 2014–2018. Wellington: reaching goals of sustainable tourism development. Museum of New Zealand Te Papa Tongarewa. Journal of Sustainable Management 11: 476–497. Accessible online at www.tepapa.govt.nz/sites/default/ Wegener, D.T., Petty, R.E. and Smith, S.M. (1995). files/statement_of_intent_2014-18_0.pdf. 24 pp. Positive mood can increase or decrease message scrutiny: Nelson, K.G. (2011). From analog prototype to digital the hedonic contingency view of mood and message drawing in the gallery. Journal of Museum Education 36: processing. Journal of Personality and Social Psychology 269–278. 69: 5–15. Packer, J. and Ballantyne, R. (2002). Motivational factors Weiler, B. and Smith, L. (2009). Does more interpretation and the visitor experience: a comparison of three sites lead to greater outcomes? An assessment of the impacts Curator 45: 183–198. of multiple layers of interpretation in a zoo context. Patel, M., Heath, C., Luff, P., vom Lehn, D. and Cleverly, Journal of Sustainable Tourism 17: 91–105. J. (2016). Playing with words: creativity and interaction in museums and galleries. Museum Management and Unpublished sources Curatorship 31: 69–86. Petty, R.E. and Cacioppo, J.T. (1996). Attitudes and Greco, A.M. (2014). Participatory exhibition design: persuasion: classic and contemporary approaches. memory jars at the Santa Cruz Museum of Art and Boulder, CO: Westview Press. 336 pp. History. Masters of Liberal Arts thesis, Harvard Petty, R.E. and Wegener, D.T. (1999). The elaboration University, Cambridge, MA. likelihood model: current status and controversies. Pp. 41–72. In: Chaiken, S. and Trope, Y. (eds). Dual process theories in social psychology. New York, NY: Guilford. 657 pp. Visitor engagement participatory activity 173

Appendix 1: Questionnaire survey

1 Did you visit Te Papa specifically to see this contemporary exhibit? (yes/no) What was your favourite aspect of this contemporary exhibit? 2 How would you rate your interest in art? (1 to 7; 1 = not interested in art; 7 = highly interested in art) 3 How would you rate your knowledge of art? (1 to 7; 1 = little knowledge of art; 7 = significant knowledge of art)

Overall, the contemporary art exhibit I attended today

4 Was enjoyable/unenjoyable (1 to 7) 5 Was meaningless/meaningful (1 to 7) 6 Did not make me curious/made me curious (1 to 7) 7 Impacted my view of my own life/did not impact my view of my own life (1 to 7) 8 Was bad/was good (1 to 7) 9 Was relevant to me/was not relevant to me (1 to 7) 10 Did not make me think/did make me think (1 to 7) 11 Was satisfying/was not satisfying (1 to 7) 12 Was not connected to anything I care about/was connected to things I care about (1 to 7) 13 Made me want to talk about what I saw/did not make me want to talk about what I saw (1 to 7) 14 Did impact my view of today’s society/impacted my view of today’s society (1 to 7) 15 Was boring/was not boring (1 to 7) 16 Was connected to things I know about/ was not connected to things I know about (1 to 7) 17 Did not make me want to know more/made me want to know more (1 to 7) 18 Impacted how I see some things in the world today/did not impact how I see some things in the world today (1 to 7) 19 Intrigued me/did not intrigue me (1 to 7) 20 Did not impact how I see myself/impacted how I see myself (1 to 7) 21 What three words would you use to describe this exhibit? 22 Did you speak with a host while in the contemporary art exhibit? If yes, what did you talk about? 23 Has your view of contemporary art changed because of the exhibit? (yes/no) 24 Did you participate in the sticky note activity? (yes/no). Why or why not? If you participated in the sticky note activity, tell us your opinion of this activity. 25 How many times (not counting today) have you visited Te Papa in the last 12 months? 26 How many were your group, including yourself? 27 What is your age? 28 Where do you usually live? (Wellington city/region, rest of New Zealand, overseas)

Tuhinga 29: 174–187 Copyright © Museum of New Zealand Te Papa Tongarewa (2018)

Evaluation of Contraception: Uncovering the collection of Dame Margaret Sparrow Stephanie Gibson

Museum of New Zealand Te Papa Tongarewa, PO Box 467, Wellington, New Zealand ([email protected])

ABSTRACT: This paper explores the context, development and evaluation of Contraception: Uncovering the collection of Dame Margaret Sparrow, an exhibition at the Museum of New Zealand Te Papa Tongarewa (Te Papa) in 2015–16. Collections and exhibitions of contraceptive material are rare in New Zealand museums. In the case of Te Papa, strategic acquisitions since 2004 have enabled the display of contraceptive objects, culminating in the stand-alone exhibition Contraception, based on Dame Margaret’s extensive collection. To guide the exhibition’s development, formative evaluation was conducted by members of the exhibition team, including the curator/author. Having exhibition staff talk directly to visitors enabled immediate understanding of our audiences, and ensured that staff could confidently champion the findings at a senior approval level and feed the results directly into the exhibition’s development. Summative evaluation followed in order to understand the impacts of the exhibition. The paper details the mechanics and findings of both evaluations, and the exhibition’s successes. It also acknowledges that curatorial assumptions regarding visitors’ perceptions of the exhibition were largely counteracted by the results.

KEYWORDS: Contraception exhibition, birth control, Margaret Sparrow, visitor research, formative evaluation, summative evaluation, contraceptive objects, sex and sexuality education.

Introduction reform. In 2002, she was made a Distinguished Companion of the New Zealand Order of Merit (later Dame Companion) Contraception: Uncovering the collection of Dame Margaret for services to medicine and the community. Te Papa’s Sparrow was an exhibition at the Museum of New Zealand exhibition honoured Dame Margaret’s contributions and Te Papa Tongarewa (Te Papa) in 2015–16.1 It was the first presented a history of contraception in a New Zealand time a substantial stand-alone exhibition had been devoted context, based on her extensive collection of contraceptive to the subject of contraception in a New Zealand museum, devices gifted to Te Papa in 2011. and it was well received by visitors, the media, historians The exhibition’s development was guided by the and the museum profession.2 findings of formative evaluation conducted by members Dame Margaret Sparrow (b. 1935) is one of New of the exhibition team, including the author/curator. This Zealand’s leading sexual-health doctors and birth- approach enabled a first-hand transformative experience control advocates. In the 1960s and 1970s, she pioneered in terms of understanding the complexity of our visitors, contraception and abortion services for students, and and in realising the power of data in terms of institutional vasectomies for men. She was one of the first doctors to approval and concept development. prescribe the emergency contraceptive pill in New Zealand, The formative evaluation was valuable for showing and has long been an outspoken advocate of abortion law that visitors favoured an approach based on liberation and Evaluation of Contraception 175 freedom, with Dame Margaret Sparrow’s integrated Collections and exhibitions into the exhibition. The majority of interviewees favoured a serious tone to the exhibition, and the majority indicated featuring contraception they would participate in writing comments within the exhibition. For those working and writing in the fields of family The cycle of evaluation was completed by conducting planning and sexual health, displays of contraceptive material summative interviews in order to understand the impacts will be familiar, from educational displays to curious devices of the exhibition, and by analysing the visitor comments. of the past (e.g. Smyth 2000: 26–27). In 1925, pioneering The findings revealed that most visitors did not know birth-control advocate Marie Stopes displayed a curio about the exhibition beforehand, but were highly satisfied cabinet of contraceptive devices in her Mothers’ Clinic in once they had visited. The exhibition generally did not central London. She called it the ‘First and Only Museum change people’s thinking about contraception, but a third of Contraception’ (Cohen 1993: 99). A descendant of this of interviewees learned more than they already knew. can be found today in Family Planning New Zealand’s Most interviewees singled out the objects as being the most head office in Wellington, where a cabinet of about 250 memorable elements of the exhibition, and most indicated contraceptive devices donated by Dame Margaret Sparrow that they would like to see similar exhibitions at Te Papa can be viewed by appointment. This semi-public display in the future. The majority of the comments written by is generally seen only by family planning staff and their visitors demonstrated constructive engagement with the visitors. exhibition’s topics. However, a search for publicly accessible museum These findings largely counteracted curatorial exhibitions devoted to contraception and associated visitor assumptions during the project that visitors would have heard research yields few examples. Notable exceptions include about the exhibition beforehand and made a purposeful the Museum of Contraception and Abortion in Vienna, visit. The exhibition team also assumed that people would Austria, which appears to be the only museum permanently be surprised, impressed or shocked by the exhibition, dedicated to the subject, and which was considered but most interviewees were familiar with the objects and groundbreaking when it opened in 2007. content. In addition, the team assumed that visitors would In 1995, the Powerhouse Museum in Sydney, Australia, find the small, humble objects inconsequential, but in fact developed a significant exhibition, Taking Precautions: the objects were identified as the most memorable parts of The Story of Contraception, from its comprehensive the exhibition. One curatorial assumption that was upheld contraceptive collection, which then toured the country. was that most visitors agreed that such exhibitions are More recently, in 2015–16, the museum displayed 190 important. contraceptive objects within the exhibition Recollect: Health That said, contraception objects are rare in most and medicine.4 museum collections and exhibitions. This may be due to Displays on contraception can be found within museums the legacies of historical attitudes, laws and social mores, dedicated to medicine and health. A significant example is which may have led museums to avoid or demote the the permanent exhibition Virtue, Vice, and Contraband: A material culture of sexuality and birth control until the late history of contraception in America, which opened in 2009 twentieth century (Brookes et al. 2013: 210; Frost 2013: in the Dittrick Medical History Center at Case Western 16; Tyburczy 2016: 101). Te Papa itself began to collect Reserve University, Cleveland (Berger 2010).5 contraceptive devices strategically only in 2004. Some museums have included contraception within Audience research on exhibitions about contraception wider contexts, such as science and technology. For is even rarer. Critique and reviews on museum exhibitions example, the Smithsonian’s National Museum of American about sex, sexuality and/or contraception can be found History in Washington, DC, displayed ‘Better than nature: (e.g. Simon 2005; Berger 2010; Frost 2013; Gosselin 2014, the pill’ as part of the exhibition Science in American Life 2015; Wannan 2015; Tyburczy 2016; Gibson 2017), but (1994–2011). An assessment of the exhibition found that there is a paucity of visitor research studies on exhibitions the majority of visitors had strongly positive views on that focus on contraception.3 This paper considers the field science, which were reinforced by the exhibition (Pekarik and contributes a case study in order to demonstrate that et al. 1995). such exhibitions can be successful and relevant to museum Two British exhibitions about sex and sexuality that audiences. included historical contraceptive material and featured 176 Tuhinga, Number 29 (2018)

visitor feedback systems were Institute of Sexology at the Contraceptive collections can also be found in Wellcome Collection, London, in 2014–15, and Intimate two of New Zealand’s medical museums (viewable by Worlds: Exploring Sexuality through the Sir Henry Wellcome appointment). The Medical Museum in Palmerston Collection at the Royal Albert Memorial Museum, Exeter, North holds several objects (also gifted by Dame Margaret in 2014. Initial analysis suggests that visitors welcomed the Sparrow), and the Cotter Medical History Museum in opportunity to think through, and express, a wide range of Christchurch holds a small collection of intrauterine ideas about sex (Jen Grove, pers. comm., 15 January 2018). devices. Young people in particular found educational value in the In general, however, most New Zealand museums objects displayed in Intimate Worlds (Fisher et al. 2017: 45). hold very little contraceptive material in their collections. The Canadian exhibition Sex Talk in the City, held A collection survey found that most institutions held only at the Museum of Vancouver in 2013, featured a range one to three contraceptive devices and/or related birth- of contraceptive material, some of it borrowed from control ephemera, although this does not account for the Dittrick Medical History Center. A key idea of the contraceptive objects that museums might not know about exhibition was that sexuality shapes people, objects and in their collections.6 As Stuart Frost notes, ‘historically, cities (Gosselin 2014: 108). Visitor research was conducted museums have found sexuality difficult to address’ (2013: to examine how visitors responded to the exhibition and 16). Objects relating to sex were either not collected or to evaluate its exhibition. Results included a high level of lay unregistered in storage, or were catalogued without satisfaction and an appreciation of the role the museum took sexual or contraceptive references. For example, museum in creating a ‘safe place’ to talk about sexuality (Gosselin collections may hold douches, but because of their multiple 2014: 113). uses, they are not generally catalogued as contraceptive In New Zealand museums, displays of contraceptive objects.7 Indeed, as observed by Jennifer Tyburczy, the devices and associated visitor research are rare. Apart from ‘raw materials’ for displaying sexual histories ‘lie dormant a long-term display of condoms at the National Army in museum archives’, therefore ultimately ‘any museum Museum in Waiouru (in the context of venereal disease can be a sex museum’ (2016: 207). during the First and Second World Wars), such displays Te Papa began to collect contraceptive material are generally short term. For example, Dame Margaret strategically only in 2004, when a large social history Sparrow and her collection were profiled in ‘The reference collection was acquired from Wellington woman Joyce collection’ in To Have and To Hold: Making Collections at Megget that included family planning literature, barrier Objectspace, Auckland (2009), which aimed to demonstrate methods and contraceptive pills. This was followed by that collectors are practitioners, working alongside artists, Dame Margaret Sparrow’s gift of the majority of her craftspeople, curators and writers (Clarke 2009: 5). Dame collection to Te Papa in 2011 (more than 750 objects). Margaret and her collection were again profiled in ‘Doctor Her collection was acquired because it related to significant to dame’ in Te Upoko o te Ika a Mäui, Wellington Museum events in New Zealand’s history (e.g. the introduction of (2013–14), an exhibition that profiled unexpected, unusual the oral contraceptive pill in 1961), and it reflected the and unique stories from the wider Wellington region. Staff life and achievements of an iconic New Zealander while observed that young adults were ‘especially interested in the also highlighting aspects of everyday life (e.g. family “Doctor to dame” story and several expressed surprise that planning) (Museum of New Zealand Te Papa Tongarewa contraception was discussed as a topic in a museum’ (Paul 2010). Since then, Te Papa has proactively collected Thompson, Wellington Museums Trust, pers. comm., 8 contemporary contraceptive technologies and filled in January 2018). significant gaps (e.g. vasectomy instruments). These Te Papa has twice positioned small displays of objects are permanently accessible at Te Papa’s Collections contraceptive devices within larger social history exhibitions Online.8 exploring particular periods of time: ‘The liberating pill?’ in These acquisitions have enabled Te Papa to create the Out on the Street: New Zealand in the 1970s (2004–05); displays mentioned above, as well as supporting external and ‘Women’s work: contraceptive revolution’ in Slice of loan requests. The lack of representation of contraception Heaven: 20th century Aotearoa (2010–17). A summative in New Zealand museums was not a factor in the evaluation of Slice of Heaven found that the ‘Women’s development of the Te Papa Contraception exhibition, but work’ objects were popular with visitors (Morris Hargreaves its rarity did make it stand out, as partly evidenced by its McIntyre 2011: 46–47). selection for the 2016 New Zealand Museum Awards. Evaluation of Contraception 177

Development of Dame Margaret The exhibition did not have an educational remit beyond these key messages, partly due to sexuality education being Sparrow’s collection and exhibition a narrow part of New Zealand’s school curriculum. That Dame Margaret Sparrow assiduously collected contraceptive said, the collection provides great potential for historical material culture throughout her career. Initially, she study, both at secondary and tertiary levels, and for sex and inherited objects from a previous director of Wellington sexuality education within the ambit of Te Papa’s vision to Family Planning Clinic, transforming this into a reference ‘be a safe place for challenging conversations’ (Museum of collection that could both illustrate the story of family New Zealand Te Papa Tongarewa 2017: 11).9 planning in New Zealand and be handled by trainee doctors and nurses (Clarke 2009: 37). Dame Margaret recalled: the bulk of it arose through my teaching and training Formative evaluation because I did a lot of work with our own staff, but also The exhibition team decided to test the following ideas providing courses for nurses, and was involved in with visitors prior to further development: medical student training, and there it was helpful to have the history of contraception as well. And it just 1 Approach: whether to focus on Dame Margaret sort of accumulated and then people were aware I was Sparrow, the science of contraception or women’s interested in objects and they would send me things liberation. as well – that’s how it grew. (Museum of New Zealand 2 Tone test: either serious or less serious. Te Papa Tongarewa 2013) 3 Visitors’ likelihood to participate in writing comments When the majority of Dame Margaret’s collection came into within the exhibition. Te Papa, curatorial staff immediately saw the potential to broaden its access beyond the health and education sectors. 4 General themes that would interest visitors. When an opportunity arose, history curators (including the author) pitched exhibition ideas about Dame Margaret and We conducted the evaluation in the main foyer of Te Papa her collection to the museum’s leadership and programming in late January 2015. Twenty visitors participated over teams. Part of our pitch was the connection between a small two days. They were solicited as they passed through the gallery within Te Papa called the Ilott Room, and the John foyer without initially being told about the content of the Ilott Charitable Trust, which had supported the gallery in the past. One of the trust’s missions is to support projects proposed exhibition. Each interview took 10–20 minutes. advancing sexual education. This particular mission had not informed museum projects before, but it contributed Results to senior management’s decision to develop an exhibition ‘positive messaging; personally interested and what on contraception. The decision was also based on the [is] happening now.’; ‘with subject matter, it [serious significance of the collection, the importance of the topic, 10 and the stature of Dame Margaret Sparrow in New Zealand. tone] sits better. Don’t make light of it.’ The exhibition team then determined the following key messages for the exhibition: To determine the approach of the exhibition, the team 1 Contraception is one of the most important inventions created show cards for each of the three possible approaches: in world history, enabling the human race to control ‘Dame Margaret Sparrow’, ‘Science of contraception’ and fertility and separate sex from reproduction. ‘Women’s liberation’ (Fig. 1). Visitors were asked which 2 Dame Margaret Sparrow is one of New Zealand’s approach they liked and why. Sixty per cent of respondents leading specialists and advocates on birth control and wanted the approach and emphasis of the exhibition to be sexual health. on liberation and freedom. About half of all respondents 3 Dame Margaret’s life, career and collection of favoured mixing two approaches, with 41% wanting contraceptive material cover a period of profound Dame Margaret Sparrow’s story to be integrated into the changes in society and technology, including the sexual exhibition, but not to be the main topic (E. MacDonald revolution (Museum of New Zealand Te Papa Tongarewa 2015a). 2015).

178 Tuhinga, Number 29 (2018)

Fig. 1 Example of the formative evaluation show cards to determine the approach of the exhibition (Gibson 2015a).

Figs 2 and 3 Examples of the tone test for the exhibition label text (Gibson 2015b). Evaluation of Contraception 179

To determine the exhibition’s tone, three images were The power of data tested with alternative label text in serious and less serious language (Figs 2 and 3). Visitors were asked which tone they The findings of the formative evaluation were unequivocal liked and why. Eighty-three per cent of respondents believed about the seriousness of the topic, and the desired emphasis the tone of the exhibition should be serious. While some on stories of liberation and freedom. respondents recognised the power of humour, there was Having exhibition team members plan, conduct and an overwhelming consensus that the topic was serious and analyse our own visitor research created a sense of ownership needed to have a serious tone (E. MacDonald 2015). in these findings (Davidson 2015: 514). In terms of ‘research To test visitors’ likelihood of participating in writing utility’, the findings had an immediate impact on the project, comments within the exhibition, two mock questions were because we could directly translate the insights into practice read to interviewees, who were asked which one (if any) they would respond to within an exhibition: (Reussner 2008: 193, 195). Pekarik (2007: 133) notes this process of transformation: ‘Team members are able to see 1. How did you first find out about contraception? in-person and up-close how visitors do or do not conform 2. How has your life changed as a result of contraception to the images that they carry of them … team members are compared to that of your parents? often effortlessly transformed in their thinking. Their minds naturally incorporate the new data of their direct experience Eighty-three per cent of respondents indicated that they would participate by answering a question, with a slightly and produce a new, expanded image of the visitor.’ higher response to the first question (E. MacDonald 2015). The research also gave us confidence. Having data – no To identify underlying themes that would be of interest, matter how small the sample – is powerful when making respondents were given a pile of 60 relevant words and decisions and seeking approval (Adams 2012: 30–31). pictures and asked to sort them into ‘interesting’ and Our results convinced Te Papa’s senior management and ‘not interesting’ piles. They were then asked to take the marketing and communications teams that the exhibition ‘interesting’ pile and sort it into thematic groups. The would be serious in tone and would integrate Dame themes of most interest were safe sex and reproduction. Margaret Sparrow’s story into the wider context of birth Eighty-three per cent of respondents indicated that they control and the sexual revolution in New Zealand society. would go to the exhibition with friends, rather than with The results directed the project’s subsequent development 11 their family or spouse (E. MacDonald 2015). and helped us write our objectives as follows:

Online survey The exhibition will profile Dame Margaret Sparrow’s Around the same time as the formative evaluation, a survey contraceptive collection, whereby visitors will was posted online to gain feedback from digital visitors, appreciate the diversity of technologies and products and to increase the number and range of respondents. and contemplate and discuss the relevance and Respondents were asked to mark a set of images and words significance of birth control to their own lives and as ‘interesting’ or ‘not so interesting’ in the context of the others, today and throughout human history. question ‘What does contraception in New Zealand mean The exhibition … will provide opportunities to to you?’ The survey was shared via Te Papa’s social media share and discuss both the personal and the universal channels (Twitter, Facebook). Of the 217 people who took journeys of women to control their own fertility. part, the majority were women (86.6%). High interest was expressed in every aspect. The lowest interest was in an (Museum of New Zealand Te Papa Tongarewa image of a condom (72%), and the highest interest (94%) was 2015b) in the phrases ‘sexual revolution’ and ‘women’s liberation’.12 180 Tuhinga, Number 29 (2018)

mainly small objects were displayed within these windows against circular background boards covered in soft wine-coloured fabric (Figs 6–9). No label text appeared alongside the objects; instead, visitors were encouraged to explore the drawers beneath each window for information (Fig. 10). A few objects were also embedded within these drawers (namely a fertility calculator, a female condom and emergency contraception).Most of the exhibition text was serious in tone, although some labels were headlined with ‘lighter’ titles, for example the introductory label for the condom display: Get it on: Condoms Today, no form of birth control is as widespread as the Fig. 4 (left) ‘The pill that changed the world’ (photo: condom – or as essential for sexual health. But it wasn’t Kate Whitley; MA_I.362295, Te Papa). always so. Condoms were often associated with promiscuity and venereal disease, and were sold The exhibition discreetly, in plain packaging. They were also notoriously unreliable. The exhibition opened on 29 May 2015, and immediately received positive press, coverage and social media Opposite the showcase, visitors could learn about Dame interest, particularly in relation to Dame Margaret Sparrow Margaret Sparrow through a brief biography and two short and her life story (N. Macdonald 2015; Mulligan 2015; audiovisuals. In the latter, Dame Margaret talked about her Wannan 2015).13 A wide range of age groups was observed life, her heroines, and significant moments in the history visiting the exhibition, from teenagers, to middle-aged of contraception and sexual health, including the impact of couples, to senior citizens.14 No negative feedback was HIV and AIDS (Fig. 11).15 received about the exhibition’s topics. On the far wall was a participatory experience titled A spotlit case formed the centrepiece of the gallery, ‘Let’s talk about sex’, where visitors were invited to answer where a single contraceptive pill was theatrically displayed the question: ‘If you could give your younger self one with the statement ‘The pill that changed the world’ piece of advice about contraception, what would it be?’ (Fig. 4). The question had become more personal and reflective One of the aims of the exhibition experience was to than originally intended, as we felt that an introspective intrigue visitors with a multitude of contraceptive devices. approach – advising oneself – would be easier to answer These included examples of early technologies used by and was more likely to engage visitors in conversation and/ people to control fertility before the pill, a wide range of or participation. condoms and marketing strategies, examples of the vast Visitors wrote their answers on circular pieces of paper, array of oral contraceptive pills manufactured since 1961, attaching them to a large graphic printed in the shape of a and modern intrauterine devices and long-acting reversible 1960s pill packet (Figs 12 and 13). and permanent contraceptive technologies. The process of choosing which objects to display was carried out by the curator and designer together, to ensure that the best objects were selected in terms of historical and technological importance, diversity of contraceptive methods for both men and women, and aesthetic qualities. A long showcase dominated one wall and was divided into four themes by circular window masks, these focusing on early birth control, condoms, the contraceptive pill and Fig. 5 The overall display of contraceptive objects long-acting contraceptive methods (Fig. 5). About 150 (photo: Kate Whitley; MA_I.362302, Te Papa). Evaluation of Contraception 181

Fig. 10 Visitors opened the drawers underneath the displays for information (photo: Kate Whitley; Te Papa, MA_I.362298).

Figs 6–9 The four themes of the exhibition: early birth control, condoms, the contraceptive pill and long-acting contraceptive methods (photos: Kate Whitley; Te Papa, MA_I.374713, MA_I.374714, MA_I.374715, MA_I.374716). 182 Tuhinga, Number 29 (2018)

Fig. 11 Audiovisuals featuring Dame Margaret Sparrow (photo: Kate Whitley; Te Papa MA_I.362309).

Summative evaluation After the exhibition opened, we were eager to complete the cycle of evaluation to see if our goals had been met, and to determine whether the exhibition had made an impact on visitors, and what those impacts might be. The latter aim was of particular interest as Te Papa headed into a major exhibition renewal project. The desire for data to both validate our approach and to inform future programming was part of our agenda. The author conducted interviews outside the exhibition over two weeks in September and October 2015. Potential respondents were solicited as they left the exhibition. Thirty- Figs 12 and 13 ‘Let’s talk about sex’ participatory activity five visitors participated in the evaluation.16 Interviews lasted and visitors’ responses (photos: Kate Whitley; Te Papa, from five to 15 minutes, capturing immediate responses as MA_I.362316, MA_I.362306). well as more in-depth reflections.17 Evaluation of Contraception 183

Results This high level of engagement can also be found in the analysis of the visitor comments left behind in the ‘Don’t see contraception in a museum every day!’18 participatory experience ‘Let’s talk about sex’. More than 2200 comments written on circles of paper survived at the The results of the summative evaluation showed that only end of the exhibition and were analysed by the author. The 20% of respondents knew about the exhibition beforehand key finding was that the majority of comments (77.5%) and had made it a purposeful part of their visit to Te indicated positive and/or constructive engagement with Papa. The majority of respondents (71%) had either seen the topics of contraception, sex, sexuality and sexual health an exhibition sign once they were inside the museum, or (Gibson 2017). This activity also provided an insight into had simply come across the exhibition as they passed by how some visitors thought about the objects and topics on (Gibson 2015c). display (e.g. ‘chocolate-flavoured condoms are a bit weird’, Generally, the exhibition did not change interviewees’ ‘100 years ago, sex was DANGEROUS’)21 (Gibson 2017). thinking about contraception. However, a third of respondents learned more about particular objects and contraceptive history: ‘Even with my background Assumptions versus findings [midwifery] I’ve seen things I didn’t know about’. ‘the diversity within the audience is always greater Sixty per cent of visitors were highly satisfied with than we can consciously accommodate’. the exhibition. A few interviewees identified limitations of the exhibition and commented on how it could have (Pekarik 2007: 134) been better. One suggested that ‘there’s room for the latest issues, e.g. young women struggling; there’s more As curator of the exhibition, I had made several assumptions to it than just contraception. We need to school up young about the people who would visit Contraception. It is very women … Abortion is worth investigating’.19 hard for a museum professional to let go of the image of Most of the interviewees (86%) identified the objects a visitor they carry in their head, and ‘often this image is as being the most memorable parts of the exhibition. Half very like the person doing the visualizing’ (Pekarik 2007: of all respondents were particularly intrigued by the early 133). Several of my assumptions about our visitors were not technology, including the douches, washable condom, sea upheld by the findings of the summative evaluation. sponge and Coca-Cola bottle top (‘Coke bottle top! Cool For example, I had assumed that people would little object … excellent storytelling device’).20 purposefully visit the exhibition, but most visitors simply Forty per cent of the interviewees knew about Dame came across it or saw a sign nearby. This finding indicates Margaret Sparrow beforehand, with just under a third that signage is critical, particularly for small exhibitions identifying her films as memorable parts of the exhibition. within an institution the size of Te Papa. One respondent observed that it was ‘great to see a woman I had assumed that people would be surprised, and handling the topic … I think those films are pretty cool – perhaps even shocked or offended by the exhibition, but it’s her talking … So valuable having women speaking in most interviewees were familiar with the topics and many context of history’. of the objects. Their pre-existing knowledge, attitudes and The majority of respondents (74%) indicated that they personal experiences about contraception appeared to be would like to see similar exhibitions about social issues at largely confirmed by the exhibition (Doering 1999: 81). Te Papa in the future, and that it was ‘highly likely’ I had thought that visitors would be impressed by Te they would visit these (71%) (Gibson 2015c). The most Papa’s courage in mounting such an exhibition. But I was commonly mentioned topics were health (including mental soon cut down to size when a visitor commented: ‘[it] health), history and change, environmental issues, and could be a lot better … It’s just a bunch of contraceptives human rights (including women’s issues and racism). One on display … When unsuccessful, it wrecks lives.’ visitor noted: ‘It’s easy to detach yourself with exhibitions I was concerned that some of the objects would be in the past … it’s easy to judge people then. Racism and dismissed as inconsequential, particularly the many small sexism are still huge. Exhibitions about important topics cardboard boxes on display. However, most visitors make things better’. identified objects as the most memorable parts of the 184 Tuhinga, Number 29 (2018)

exhibition. This was partly due to successful design strategies visitation was generally casual and therefore signage was (e.g. the arrangement of the contraceptive pill packets from critical in attracting people’s attention, and they were A to Z),22 and to the inclusion of both familiar objects and humbling in that most visitors came with prior knowledge unexpected items such as the Coca-Cola bottle top. and their own narratives. The findings were reassuring in An assumption that was upheld by the findings was that that the most memorable elements of the exhibition were most visitors agreed that exhibitions like Contraception are the objects, regardless of how small or ordinary, and that the important. I received many emphatic answers of ‘absolutely’ majority of interviewees wanted to see similar exhibitions at and ‘definitely’ when visitors were asked if they would like Te Papa in the future. to see more such topics displayed at Te Papa.

Conclusion Acknowledgements Many thanks go to Dame Margaret Sparrow for her From a curatorial perspective, Contraception: Uncovering the generous advice and support throughout the project. I collection of Dame Margaret Sparrow was a success because thank Dr Edith MacDonald (former Head of Insights, Te it met our objectives to profile Dame Margaret Sparrow’s Papa) for encouraging me to undertake and publish the collection and enable visitors to appreciate the diversity research. I thank Dr Jen Grove (University of Exeter, of objects, contemplate the relevance and significance United Kingdom) and Viviane Gosselin (Museum of of birth control, and have the opportunity to share and Vancouver) for sharing their work and insights, and Tilly discuss their own journeys in controlling fertility. It also Boleyn (Museum of Applied Arts and Sciences, Sydney) brought a nationally significant collection into wider public for her support. I thank Te Papa’s Liaison Librarians for consciousness than in its previous context as a semi-public providing resources. I also thank the reviewers for their sexual health and education collection. excellent guidance in revising the paper. The development of the project was also successful because it was informed by the findings of visitor research. Researching our visitors through formative and summative evaluations gave us better tools and confidence to create the exhibition, and to take our findings into future programming and practice. In addition, having members of the exhibition team conduct their own research was a transformative experience, both in terms of understanding the complexity of visitors and challenging our assumptions, and in realising the power of the data to gain institutional approval. The formative evaluation was valuable for showing that visitors favoured an approach based on liberation and freedom, with Dame Margaret Sparrow’s story integrated into the exhibition. The majority of visitors favoured a serious tone to the exhibition, and indicated they would participate in writing comments within the exhibition. The 2200-plus comments written by visitors indicated a high level of constructive engagement with the exhibition’s themes. The summative evaluation was valuable for showing that topics such as contraception are of interest to our audiences. The key findings were, in turn, prosaic, humbling and reassuring. They were prosaic in that Evaluation of Contraception 185

Notes at Te Papa in 2012 with the simple classification of ‘douches’. It may have been unregistered in storage 1 On display in the Ilott Room, Level 4, Te Papa, 29 for decades. Its catalogue record (registration number May 2015 to 31 January 2016. GH024681) has been upgraded to reference ‘sex’ and 2 The exhibition was a finalist in the social history ‘birth control’ (Museum of New Zealand Te Papa category of the New Zealand Museum Awards in Tongarewa n.d.). 2016 (Museums Aotearoa 2016: 8). In an issue of the 8 Te Papa’s Collections Online can be accessed at New Zealand Journal of History devoted to New https://collections.tepapa.govt.nz. Zealand’s sexual histories, Chris Brickell and Angela 9 A successful model of practice is the University of Wanhalla observed that museums are taking the lead in Exeter’s Sex and History project, which uses objects contributing sexual histories drawn from sources other from the past to talk about sex and sexuality with young than writing, citing the material culture of Contraception: people (Fisher et al. 2017). Uncovering the collection of Dame Margaret Sparrow as their example (Brickell & Wanhalla 2016: 9). 10 Interviewee quotes from the formative evaluation (E. MacDonald 2015). 3 Internal reports may reveal such data (e.g. Pekarik et al. 1995; Hall 2013). Direct communication with 11 This particular finding played out in the summative museum staff can also tease out information on evaluation, where visiting with friends dominated audiences (see Note 5). (Gibson 2015c). 4 Taking Precautions was curated by Megan Hicks, 12 Online survey results that indicated the degree of who had collected contraceptive material in the 1980s interest in selected images and phrases: image of a when it was considered ‘quite daring’. Recollect: Health condom (72.1%); image of a Family Planning poster and Medicine (August 2015 to mid-2016) was curated (73.3%); the word ‘condom’ (75.7%); ‘morality’ by Tilly Boleyn, who noted that the exhibition ‘sparks (76.4%); image of contraceptive pills (81.1%); ‘health a lot of discussion between people, particularly women’ scares’ (83.2%); image of douche (83.9%); image of (pers. comm., 18 January 2016). female condom (86.2%); image of women’s health 5 The Virtue, Vice, & Contraband exhibition was protest (86.8%); the word ‘pleasure’ (90.9%); image drawn from a large contraceptive collection donated of a Dalkon Shield IUD (91.1%); ‘invention’ (93.2%); by Canadian Percy Skuy, a past president of the Ortho the phrases ‘sexual revolution’ (94.1%) and ‘women’s Pharmaceutical company in Toronto (Case Western liberation’ (94.6%). Respondents were 86.6% female Reserve University 2010). Skuy’s collection had and 11.1% male (2.3% preferred not to say). formerly been on display in the company’s headquarters 13 Radio coverage included Kim Hill interviewing Dame since the 1960s. In terms of audience research, James Margaret Sparrow on Radio New Zealand, on 6 June Edmondson, Chief Curator of the Dittrick Medical 2015 (Hill 2015), and Teresa Cowie interviewing the History Center, noted that Virtue, Vice, & Contraband curator/author on Radio New Zealand, on 12 January had ‘been very well received’ and had been a ‘real 2016. game changer for us, in an overwhelmingly positive way. It has pivoted us from being a doctors’ museum 14 Observed by Te Papa Visitor Host Basil Chan (pers. to embrace the patient’s experience. Demographics comm., 16 September 2015). on our museum Facebook page show that women 15 The audiovisuals can be seen on YouTube under the 25–35 make up over 60% of our traffic there’ (James titles ‘Birth control: changing womans lives’ (Te Papa) Edmondson, pers. comm., 29 October 2015). and ‘From condoms to consumer rights’ (Te Papa). 6 The author surveyed 17 museums throughout New The second of these begins with an image of Dame Zealand in late 2015 about their contraceptive Margaret Sparrow in front of her display cabinet at holdings and whether such material was on display. Family Planning’s head office in Wellington.

7 Pahiatua and Districts Museum Society holds a douche 16 Summative evaluation demographic breakdown: gender can that is classified as a ‘household object’ (eHive – females (68.5%), males (31.5%); age – 16–30 years 2018). A similar douche can was ‘found in collection’ (34.2%), 30–50 years (37%), 50–80 years (18.4%); 186 Tuhinga, Number 29 (2018)

ethnicity – New Zealand European/Päkehä (65.7%), References other European (20%), Asian (11.4%), Pasifika (2.8%); place of origin – Wellington (17.1%), rest of New Adams, M. (2012). Museum evaluation: where have we Zealand (57.1%), international (25.7%); type of visit – been? What has changed? And where do we need to go with friends (37%), visiting alone (34.2%), with family next? Journal of Museum Education 37(2): 25–36. member (20%), with a partner (17.1%). Berger, M.W. (2010). Virtue, Vice, and Contraband: A history of contraception in America (review). Bulletin of 17 Summative evaluation questions: the History of Medicine 84(2): 281–282. 1. Did you know about Dame Margaret Sparrow Brickell, C. and Wanhalla, A. (2016). Editorial introduction: before visiting? writing New Zealand’s sexual histories. New Zealand 2. Has this exhibition changed your way of thinking Journal of History 50(1): 3–12. about contraception? How? Brookes, B., Gooder, C. and de Castro, N. (2013). ‘Feminine as her handbag, modern as her hairstyle’: the uptake of 3. What was your overall experience of the the contraceptive pill in New Zealand. New Zealand exhibition? Journal of History 47(2): 208–231. 4. Why did you visit this exhibition today? Case Western Reserve University (2010). The Percy 5. What is the most memorable part of the exhibition? Skuy collection. In: Case Western University [website]. 6. This exhibition talks about contemporary social Retrieved 20 October 2015, from www.case.edu/ issues such as women’s health. Would you like to affil/skuyhistcontraception/index.html. see similar exhibitions at Te Papa in the future? Clarke, P. (2009). To Have and To Hold: making collections. 7. What is the likelihood of visiting similar Auckland: Objectspace. 47 pp. exhibitions? Cohen, D.A. (1993). Private lives in public spaces: Marie Stopes, the Mothers’ Clinics and the practice of 18 Interviewee quotes in this section are from the contraception. History Workshop Journal 35: 95–116. summative evaluation (Gibson 2015c). Cowie, T. (2016) Te Papa exhibition showcases 19 The topic of abortion was not explored in the contraceptives collection, Summer Report, Radio New exhibition. Zealand, 12 January. Retrieved on 13 January 2016 from www.radionz.co.nz/national/programmes/ 20 A Coca-Cola bottle top was displayed to represent the summerreport/audio/201785331/te-papa-exhibition- use of the effervescent drink as a vaginal douche after showcases-contraceptives-collection. sex in the 1950s–60s, when it was difficult for young Davidson, L. (2015). Visitor studies: toward a culture of people to obtain contraception before marriage. reflective practice and critical museology for the visitor- 21 Contraception exhibition – visitor comments (originals) centred museum. In: McCarthy, C. (ed.). The – 852917 (Te Papa Corporate Records). international handbooks of museum studies: museum practice. Chichester: John Wiley & Sons, Ltd. Pp. 22 An example of the success of the A to Z arrangement 503–527. of the contraceptive pill packets can be seen in its Doering, Z.D. (1999). Strangers, guests, or clients? Visitor inclusion in artist Ann Shelton’s work Dark matter experiences in museums. Curator 42(2): 74–87. (2016: 193–195). eHive (2018). Simplex douche can; Simplex; 1978- 0580-1. In: eHive [website]. Retrieved 4 January 2015, from https://ehive.com/collections/3136/objects/602536/ simplex-douche-can. Fisher, K., Grove, J. and Langlands, R. (2017). ‘Sex and history’: talking sex with objects from the past. In: Allen, L. and Rasmussen, M.L. (eds). The Palgrave handbook of sexuality education. London: Palgrave Macmillan. Pp. 29–51. Frost, S. (2013). Museums and sexuality. Museum International 65(1–4): 16–25. Evaluation of Contraception 187

Gibson, S. (2017). ‘Let’s talk about sex’: visitor comments Tyburczy, J. (2016). Sex museums: the politics and in Contraception: uncovering the collection of Dame performance of display. Chicago and London: University Margaret Sparrow. Curator 60(1): 47–65. of Chicago Press. 286 pp. Gosselin, V. (2014). Civic museography, porous narratives Wannan, O. (2015). History of birth control laid bare. and the choir effect: Sex Talk in the City at the Museum Dominion Post, 1 June, p. A8. of Vancouver. THEMA 1: 107–115. Gosselin, V. (2015). Debunking, decentralizing and dissonance: cultural jamming @ Museum of Vancouver. In: Urtizberea, I.A. (ed.). El desafio de exponer. Bilbao: Unpublished sources Universidad de País Vasco. Pp. 47–70 . Hill, K. (2015). A NZ history of contraception. Saturday Gibson, S. (2015a). Test approach. Visitor & Market Morning, Radio New Zealand, 6 June. Retrieved Research, EP-EX-Margaret Sparrow Exhibition 17 November 2015, from www.radionz.co.nz/national/ [archival folder], Museum of New Zealand Te Papa programmes/saturday/audio/201757346/dame-margaret Tongarewa, Wellington. -sparrow-contraception-history. Gibson, S. (2015b). Test tone 1. Visitor & Market Research, Macdonald, N. (2015). Portrait of a pioneer. Sunday EP-EX-Margaret Sparrow Exhibition [archival folder], [Sunday Star-Times magazine], 6 September, pp. 14–17. Museum of New Zealand Te Papa Tongarewa, Mulligan, J. (2015). The Golden Speculum. Salient, 16 Wellington. August. Retrieved on 17 November 2015 from http:// Gibson, S. (2015c). Visitor experience evaluation report salient.org.nz/2015/08/the-golden-speculum. Contraception. Visitor & Market Research, EP-EX- Museum of New Zealand Te Papa Tongarewa (n.d.). Margaret Sparrow Exhibition [archival folder], Museum Enamel douche can. In: Collections Online [online of New Zealand Te Papa Tongarewa, Wellington. database]. Retrieved 4 January 2018, from https:// Hall, J. (2013). Visitor study report of ‘Sex Talk in the City’ collections.tepapa.govt.nz/object/1325857. exhibition. Museum of Vancouver, Vancouver. Museum of New Zealand Te Papa Tongarewa (2017). MacDonald, E. (2015). Margaret Sparrow Evaluation write- Statement of intent 2017–2021. Wellington: Museum up. GV-OS-18-01, Museum of New Zealand Te Papa of New Zealand Te Papa Tongarewa. 19 pp. Tongarewa, Wellington. Museums Aotearoa (2016). Service IQ New Zealand Morris Hargreaves McIntyre (2011). ‘What it is to be Kiwi’: Museum Awards 2016. Wellington: Museums Aotearoa. a summative evaluation of Slice of Heaven: 20th century 18 pp. Aotearoa. VM-AA-05-04-03, Museum of New Zealand Pekarik, A.J. (2007). Studying visitors and making museums Te Papa Tongarewa, Wellington. better. Curator 50(1): 131–134. Museum of New Zealand Te Papa Tongarewa (2010). Pekarik, A.J., Doering, Z.D. and Bickford, A. (1995). Acquisitions Strategy 2010/11. GV-OS-05-02, Museum An assessment of the ‘Science in American Life’ exhibition of New Zealand Te Papa Tongarewa, Wellington. at the National Museum of American History. Smithsonian Institution Report 95-5. Washington, DC: Smithsonian Museum of New Zealand Te Papa Tongarewa (2013). Institution. 102 pp. Oral history recording, Dame Margaret Sparrow. Reussner, E.M. (2008). Learning from the best: success CA001174/001/0002, Museum of New Zealand Te factors for effective audience research. In: Kräutler, H. Papa Tongarewa, Wellington. (ed.). Heritage learning matters: museums and universal Museum of New Zealand Te Papa Tongarewa (2015a). B1 heritage; proceedings of the ICOM/CECA ’07 Story and Experience (Hidden Histories: Dame Margaret Conference. Vienna: Schlebrügge. Pp. 186–197. Sparrow’s Contraceptive collection). Story & Experience Shelton, A. (2016). Dark matter. Auckland: Auckland Art Development-General, EP-EX-Margaret Sparrow Gallery. 293 pp. Exhibition [archival folder], Museum of New Zealand Simon, S. (2005). Bringing contraception history out of the Te Papa Tongarewa, Wellington. closet/Museum has unusual artifacts, from gold cervical Museum of New Zealand Te Papa Tongarewa (2015b). caps to earwax. Los Angeles Times, 2 January. B3 Exhibition Delivery Plan (Dame Margaret Sparrow’s Smyth, H. (2000). Rocking the cradle: contraception, sex and Contraceptive collection). EP-EX-Margaret Sparrow politics in New Zealand. Wellington: Steele Roberts Ltd. Exhibition [archival folder], Museum of New Zealand 253 pp. Te Papa Tongarewa, Wellington.