An Update from Papua New Guinea

Tracing Domestication and Cultivation of Bananas from Phytoliths: An update from Papua New Guinea

Carol J. Lentfer

Research

Abstract

There is now good evidence from current banana distribu- (bell), seeds and flowers can also be eaten; the leaves tions and genetic analysis that Papua New Guinea and are used for cooking, wrapping, serving food and for shel- nearby regions have played a key role in the domestica- ter from the sun and rain; fibers from the stem and peti- tion of edible Eumusa and Australimusa bananas. Strong oles are used for making string, rope and other cordage support for this also comes from phytoliths in the archaeo- for weaving; the sap is used for dye; and, the seeds are botanical record. Seeds have diagnostic phytoliths which used as beads or money (Burkill 1935, Lentfer 2003a, can be used to discriminate between the two main sec- b). Given this multitude of uses, even in their wild state tions of edible bananas, the giant banana, Musa ingens, and prior to the development of fleshy, seedless (or near- and Ensete. Therefore, the presence of seed phytoliths ly seedless) fruit (Lentfer 2003a), it is likely that banana and their subsequent disappearance from archaeological plants would have been recognized as a prized resource assemblages can be used to trace processes of domes- and exploited in the past in ways similar to the present. tication leading to parthenocarpy and sterility. Following Indeed, this is the picture that is gradually emerging from loss of viable seeds, banana presence can still be docu- a host of archaeobotanical, biogeographical, biomolecu- mented from phytolith morphotypes from other plant parts, lar and genetic evidence (e.g., Carreel et al. 2002, De particularly the volcaniform morphotypes from leaves. Langue & de Maret 1999, Denham et al. 2003, Kennedy Nevertheless, according to several pioneer studies, these 2008, Mbida et al. 2001, Perrier et al. 2009, Vrydaghs are more difficult to differentiate unless they occur in re- et al. 2009). Complex origins and multiple dispersals for gions where certain species or varieties of bananas are banana cultivars are indicated, but perhaps more strik- not endemic. ing is the significant role that the New Guinea region has played in the development of the domesticated banana. After a brief review of the evidence for the origins and This paper reviews results from morphometric and mor- spread of cultivated bananas, this paper outlines a cur- photypic analyses of Musaceae phytoliths and briefly in- rent research project which expands current banana phy- troduces the ‘New Guinea Banana Project’ which builds upon previous analyses. The morphometric database, combined with a comprehensive set of images, facilitates banana phytolith identification and is another step forward in solving the issues surrounding banana dispersal, cul- Correspondence tivation and domestication, especially in the Pacific/New Carol J. Lentfer, School of Social Science, University of Queen- Guinea region. sland, Queensland, AUSTRALIA. [email protected] Introduction

Musa bananas (including plantains) constitute major food staples in the tropics and in terms of gross value of food production they are by far the most important world fruit crop. Furthermore, the banana plant is valued for more Ethnobotany Research & Applications 7:247-270 (2009) than just its fruit (Kennedy 2009): the stem, corm, bract Published: July 30, 2009 www.ethnobotanyjournal.org/vol7/i1547-3465-07-247.pdf 248 Ethnobotany Research & Applications tolith databases for further assessment of phytolith varia- ularly phytoliths (reviewed in Denham & Donohue 2009 tion within and between wild and domesticated Musa ba- and Donohue & Denham 2009). Evidence from starch nanas and Ensete in Papua New Guinea. shows good promise of adding to this (Lentfer 2009), but currently there is only one confirmed record with a positive Origins and distributions of banana cultivars identification of banana starch associated with an archaeological deposit from Santa Cruz in the Solomon Islands There are several wild species of Musa bananas and (Crowther 2009). The only other record comes from the closely related Ensete species ranging from Africa (Ensete Yuku rock shelter site in the Western Highlands of Papua only) to India, Southeast Asia and as far east as New New Guinea, but this has not been confirmed as banana Guinea and the Solomon Islands (Ensete and Musa). Ad- (Horrocks et al. 2008). Most phytolith records to date have ditionally, there are hundreds of diploid, triploid and poly- been based on the identification of distinctive volcaniform ploid cultivars derived from a few wild species from two morphotypes from Musaceae leaves. However, identifica- sections of bananas (Eumusa and Australimusa) that are tion beyond the family level has not been successfully at- cultivated in tropical and subtropical regions worldwide tempted in most studies because of the difficulties in dis- (Argent 1976, Arnaud & Horry 1997, Daniells et al. 2001, criminating between volcaniform morphotypes from differ- De Langhe et al. 2009, Kennedy 2008, Lentfer 2003a, ent Musaceae genera, sections and species. Pollefeys et al. 2004, Sharrock 1990, Simmonds 1959, 1962, Valmayor 2001, Wong et al. 2001). Phytogeogra- In his pioneering work at Kuk in the Western Highlands phy and genetic evidence shows that the domesticated of Papua New Guinea, Wilson (1985) used morphometric Australimusa Fe`i banana almost certainly originated in analyses to discriminate between three sections of banan- the New Guinea region (Jarret et al. 1992, MacDaniels as and he also identified some Musaceae morphotypes in 1947, Simmonds 1959), but the origins of the more com- sediments dated to c. 10,000 cal B.P. as Australimusa. monly known and widely marketed Eumusa cultivars have However, these identifications were problematic and in- been more difficult to determine. Based on the distribution conclusive, partly due to the limited set of comparative ref- of diploid and triploid Eumusa bananas that contain an A erence material analyzed, but also due to the assumption genome derived from either Musa acuminata Colla sub- that Eumusa section bananas were introduced into New species banksii (F. Muell.) Simmonds or errans (Blanco) Guinea from Southeast Asia during the mid-to-late Holo- R.V. Valmayor or both (see summary in Table 6, Kennedy cene (e.g., Spriggs 1996). More recently, Lentfer (2003a) 2008:85), it appears that the primary center for the do- found that the seeds from Australimusa, Eumusa, Ingenti- mestication of edible Eumusa section bananas was the musa and Ensete have diagnostic phytolith morphotypes Philippines/New Guinea region. (Figure 1), and was able to confirm the presence ofMusa ingens Simmonds - the giant cold-tolerant banana be- Significantly, a long history of banana manipulation by hu- longing to Section Ingentimusa, Ensete glaucum (Roxb.) mans, particularly in the New Guinea region, is indicated Cheesman, and Eumusa section bananas from seed phy- by the presence of the banksii A genome in AAB plan- tolith morphotypes in a similar archaeological context at tains as far afield as Africa and the Pacific. Coupled with the Kuk Swamp site (Denham et al. 2003, Lentfer 2003b). the proliferation of diploid AA cultivars in New Guinea, this The earliest records for Eumusa and Ingentimusa seed points to the likelihood of an early dispersal from the New phytoliths recovered from palaeochannel fills in this analy- Guinea region in two directions, eastwards into the Pa- sis were dated at c. 10,000 cal B.P. Eumusa persisted to cific region and westwards through Island Southeast Asia, the top of the archaeological sequence dated at c. 2500 Malaysia and across to Africa (Kennedy 2008:85-86). The cal B.P. The earliest date for Ensete seed phytoliths re- large number of diploid AA cultivars with both banksii and covered from palaeosurface feature fills was c. 7000-6500 errans genomes, as well as the absence of the errans ge- cal B.P. Similar to Eumusa, these persisted to the top of nome in the African AAB plantains but its presence togeth- the sequence. Volcaniform leaf phytoliths were also pres- er with the banksii genome in the Pacific AABMaia Maoli ent but were not identified to any particular section or spe- plantains are interesting and imply complex species and cies. subspecies interactions within the Philippines/New Guin- ea regions. This would have involved human diffusion of In addition to the Kuk site, Lentfer also confirmed the bananas, probably concurrent with the earliest transfer of presence of Eumusa section bananas from diagnostic the banksii A genome westwards into Island Southeast seed phytoliths at the coastal Lapita site of SAC on Wa- Asia, and then to mainland Asia and eventually Africa, and tom Island, East New Britain, Papua New Guinea (Lentfer also subsequently over an extended period as people ex- & Green 2004), as well as Ensete, Eumusa and Australi- panded eastwards into the Pacific (Kennedy 2008, Perrier musa section bananas at the archaeological site of FIF/4 et al. 2009). at the Yombon airstrip in South West New Britain (Lentfer et al. 2008). As with Kuk, volcaniform leaf phytoliths were Archaeobotanical records for bananas present in both assemblages but no attempt was made to identify them beyond the family level in the initial analy- The archaeobotanical record for bananas is very sparse ses. and is mostly derived from microfossil evidence, partic-

www.ethnobotanyjournal.org/vol7/i1547-3465-07-247.pdf Lentfer - Tracing Domestication and Cultivation of Bananas from Phytoliths: 249 An update from Papua New Guinea A B

C D

E F

Figure 1. Diagnostic seed morphotypes of wild Musa bananas and Ensete from Papua New Guinea. A-B. Ensete glaucum (Accession No. QH28807); C-D. Musa ingens, Section Ingentimusa (Accession No. WH1); E-F. Musa peekelii, Section Australimusa (Accession No. WNB488).

www.ethnobotanyjournal.org/vol7/i1547-3465-07-247.pdf 250 Ethnobotany Research & Applications

G H

I J

K L

Figure 1 cont. Diagnostic seed morphotypes of wild Musa bananas and Ensete from Papua New Guinea. G-H. Musa maclayi, Section Australimusa (Accession No. MB6); I-J. Musa acuminata ssp.banksii, Section Eumusa (Accession No. QH067962); K-L. Musa schizocarpa, Section Eumusa (Accession No. NB489).

www.ethnobotanyjournal.org/vol7/i1547-3465-07-247.pdf Lentfer - Tracing Domestication and Cultivation of Bananas from Phytoliths: 251 An update from Papua New Guinea Apart from phytolith studies in New Guinea there are few conut, Canarium, Job’s tears, possibly sugar cane, phyto- other published accounts of banana identification beyond liths derived from pioneer tree species and grasses that the family level with the exception of two African studies colonise gardens, and pig and chicken bones (Lentfer & in Cameroon and Uganda. Volcaniform morphotypes re- Green 2004). All other evidence for cultivation primarily covered from refuse pits at an agricultural village site, relies on bananas being outside their natural range of dis- Nkang in southern Cameroon dating to c. 2500 cal B.P. tribution. For instance, bananas are outside their natural were identified to genus Musa (Mbida et al. 2001, 2004). range in the Pacific east of the Solomon Islands. There- Further to the east, at Munsa, Uganda, both Musa and fore, all records for bananas east of the Solomons are Ensete phytoliths were identified from swamp sediment indicative of human translocation and cultivation includ- cores (Lejju et al. 2006). The oldest dates for these were ing wild M. acuminata ssp. banksii found in Samoa (De c. 5200 cal B.P. (cf. Neumann & Hildebrand 2009). In con- Langhe 2009, MacDaniels 1947) and the Musa found as- trast to New Guinea, no seed phytoliths were recorded sociated with Lapita deposits in Vanuatu (Horrocks et al. from either African site, probably because Musa bananas in press). The same applies in Africa where only Ensete were already seedless by the time they had been intro- species are indigenous. Musa banana cultivation can be duced into Africa. inferred at the Nkang site in Cameroon by 2500 cal B.P. from the presence of Musa phytoliths (Mbida et al. 2001, The question of cultivation 2004) and pending the accuracy of dating and the morphotypic discrimination between Musa and Ensete volca- Banana plants in their natural state are light-demanding niform phytoliths (Neumann & Hildebrand 2009), it is pos- pioneer species of tropical environments. Growing nat- sible that banana cultivation occurred in Uganda as early urally from seed and via suckers, bananas are adapted as c. 5200 cal B.P. (Lejju et al. 2006). to opportunistic colonization of mostly well-drained open sites such as forest margins, forest gaps resulting from Identifying banana cultivation in areas where wild banan- tree fall, and scree slopes associated with landslides and as grow naturally is perhaps the most difficult, especially erosion (e.g., Argent 1976, MacDaniels 1947). Human in the absence of supportive archaeological and/or palae- selection, which eventually led to female sterility, loss of obotanical evidence. At the Yombon airstrip locale, south seeds and parthenocarpy, has produced hundreds of dif- West New Britain, Papua New Guinea, for instance, ba- ferent land races and hybrids of Eumusa, Australimusa nanas appear in the early to middle Holocene. Howev- and Eumusa x Australimusa section bananas occurring in er, although this coincides with a major burning episode, the Indo-Pacific region and Africa (see De Langhe et al. there is no other evidence supportive of cultivation other 2009) and implies a long history of somatic mutation and than the presence of a few potential cultivars including human manipulation involving cultivation. Saccharum sp. and Coix lachryma-jobi L. In cases such as this, cultivation could only be confirmed if it could be Nevertheless, given the sparse record of prehistoric ba- proven that bananas were indeed seedless, and therefore nana distribution, tracking evidence for cultivation and dis- cultivars. persal of cultivars is difficult. The record is derived mostly from two broad categories of evidence: ecological and Identification of cultivars: geographical. This is mostly reliant on archaeobotanical finds showing presence of bananas: outside their natural morphometric and morphotypic range; in contexts with archaeological, sedimentary and analyses of phytoliths ecofactual features indicative of cultivation; or, in association with other known domesticates and associated spe- Tracing the history of banana cultivation, domestication cies, plants and/or animals. At Kuk, for instance, the alti- and dispersal could be greatly facilitated if phytoliths can tude is exceptionally high (>1560 metres above sea lev- be readily differentiated in archaeobotanical assemblag- el) for wild Eumusa section bananas and Ensete to occur es. However, the variation of phytoliths within and be- naturally (see Argent 1976), but even if the early Holocene tween Musaceae taxa and within and between plant parts environment was warmer than it is today, the presence (with the exception of banana seed phytoliths) has, until of stake holes, post holes and mounds, coupled with the recently, been poorly understood. To address this shortfall relatively high proportions of Musaceae phytoliths in the and to expand previous morphotypic and morphometric phytolith assemblages, particularly following erosion and analyses undertaken by Wilson (1985) and Mbida et al. burning episodes, are strongly supportive of human influ- (2001), phytolith researchers have commenced a series ences and cultivation at least by about 7000-6500 cal B.P. of rigorous morphometric analyses of banana phytoliths to (Denham et al. 2003). determine their diagnostic value (e.g., Ball et al. 2006).

Evidence for cultivation at the Watom site is equally strong. Lentfer (2003a) has undertaken preliminary studies to in- Banana phytoliths are found in black, humic rich soils typi- vestigate the variation of phytoliths within and between cal of gardens. They are in association with phytoliths and species and also within and between plant parts. The first macrobotanical remains from other cultivars including co- set of exploratory analyses examined seventeen acces-

www.ethnobotanyjournal.org/vol7/i1547-3465-07-247.pdf 252 Ethnobotany Research & Applications

Table 1. Banana accessions analysed by Lentfer (2003a, b). A Section Wild species Accession code (plant parts*) Eumusa Musa acuminata QH325354 ssp. banksii (lf,sd,sk,mrb,ped); QH541190 (sd,br) Musa schizocarpa 489 (lf,sd,sk); QH356650 (lf) Australimusa Musa peekelii QH067966 (lf,sd); 488 (lf;sd/ fr); 489 (fr/sd); QH067968 (sk) Musa maclayi QH537000 (lf,br); NB487 (lf,lfbs/st); QH356648(mrb) N/A Ensete glaucum QH28807 (sd,sk); B 482 (lf,sk,sd); QH356652 (lf) Cultivars Eumusa Musa acuminata QH438477 (lf) Musa paradisiaca? QH4000037 (lf) Australimusa TT(Fe`i) QH067969 (lf) *lf=leaf; sd=seed; fr=fruit; sk=skin; br=bract; mrb=mid rib; st=stem; lfbs=leaf base; ped=peduncle Note Fe`i is referred to as M. fei F. Muell. in figures. sions consisting of Ensete and wild and cultivated Australi- musa and Eumusa Section bananas (Table 1). Twenty-five phytoliths from a number of different plant parts including leaf blades, leaf mid-ribs, leaf bases, fruit and seed, skin, pseudostems, bracts and peduncles were analyzed separately. Only phytoliths with craters (i.e., spherical to C sub-spherical to blocky morphotypes but not necessarily volcaniform morphotypes, see Figure 2) were included in the analysis. It did not include any of the clearly diagnostic seed morphotypes referred to previously and shown in Figure 1.

Analysis 1

a) Differentiation of phytolith seed morphotypes from other plant parts: Long dimensions of phytolith bodies and craters were measured (see Figure 1, Vrydaghs et al., 2009) and ratios of body length to crater width were calculated. Results of the analysis using pooled data show that the ratio of mean body length to mean crater width is significantly different at α = 0.05, differentiating Figure 2. A. Sheet of polygonal and globular seed between seed/fruit pulp morphotypes and morphotypes phytoliths from Musa acuminata ssp. banksii. These from other plant parts (Figure 3). Notably, body length morphotypes have craters and were included in the and crater width scores by themselves were less helpful analysis. The plant material was obtained from the in this regard. Queensland Herbarium (Accession No. QH067962). B-C. Examples of volcaniform and globular leaf morphotypes b) Differentiation of Eumusa, Australimusa seed mor- from Musa maclayi (Accession No. NB487) examined in photypes and Ensete: Ensete glaucum has distinctive the analysis.

www.ethnobotanyjournal.org/vol7/i1547-3465-07-247.pdf Lentfer - Tracing Domestication and Cultivation) of Bananas from Phytoliths: 253 sk ( ) An update from Papuaglaucum New Guinea sk E. 9 ( A A 8 glaucum E. 8 (lf) ) glaucum 8 6 8 sk (lf) E. 7 ( glaucum 7 glaucum 37 E. midrib E. ) 7

6 fr 6 6 / 4 ) peekelii sd 25 fr ) ( / (lf) crater dimension crater 25 crater dimension crater sk

5 M. ( peekelii glaucum sd 25 ( glaucum Cl Cl eaf/stem l E. crater dimension crater M.

4 4 E. 2 7 24 4 Cl ) 95% 95% ) eekelii sd 24 ( fr p ) ) 95% / 100 3 (lf) eekelii sd sk peekelii sd M. eed/fruit ( ( 25 p glaucum 2 2 ( s 17

0 glaucum crater dimension crater M. E.

N= M. 25 17 24 25E. 7 7

2 Cl 17 21N- 332 100 25 37 ) species M. banksii M. peekelii M. peekelii E. glaucum E. glaucum anksii ) sd 332 BB 24 ( b ) (sd) ) (sd) (sd/fr) (lf) (sk) ) 95% fr Part leaf/bract seed/fruit leaf/stem midrib 25 0 0 / eaf/bract anksii sk eekelii l sd sd (lf) ( 3A 20 M. 95% Cl crater dimension N= 4A N= (

( 24 b 24 17 17 p 25 25 25peekelii 7 25 7 sd sp. - 25 glaucum 25 glaucum 95% Cl( crater dimension ) N

9 8 7 6 5 4 3 2 M. banksii M.M. p beekeliianksii M. peekeliipeekelii E.M. glaucum peekelii E.E. glaucum glaucum E. glaucum

species species dimension crater E. N= M. 4 2 E. M. M. 8 6 0 sk Part ( 19 17 7 ) (sd) ((sdsd)) (sd(sd/fr) ) Cl (sd(lf)/fr) (sk(lf)) (sk) 37 3A N= sk 4 2 8 6 0 Ensete species midrib ( 37 ) 4A 4A 4A 24 ) 7 glaucum ) fr 95% midrib 18 B 95% Cl crater95% dimension Cl crater dimension anksii / (lf) sd E. species eekelii sd 4A ( b sp. ( glaucum peekelii p 25 sd 25 6 (

17 E. M. crater dimension crater (lf) 25 longdimension M. M. 7

8 ) 8 17 Cl 25 longdimension Cl eaf/stem N= sk 4 2 (lf) 8 6 0 Ensete l ( 16 5 ) Cl glaucum 6 eaf/stem glaucum fr 24 l ) ) / 95% 95% E. E. species anksii 4A sd ) eekelii peekelii sd sd 7 ( b 25 95% 4

15 ( ( p 25 fr 6

6 / crater dimension crater 100 M. ) M. M. peekelii sd 100 eed/fruit 25 Cl fr ) ( 17 s 14 3 / (lf) eed/fruit crater dimension crater N= sk 4 2 8 6 0 M. s ( peekelii glaucum sd 24 25 ( glaucum Cl ) 95% ) E. 13 2 4 dimension crater M. eekelii sd E. species 4 anksii 4A 7 sd N- ( 332 100 25 37 p ( b 332 22 9 Cl ) 25 C 95% 332 Part eaf/bract leaf/bract seed/fruit leaf/stem midrib ) eekelii M. sd l 1 M. ( 24 17 p fr ) - eaf/bract ) 3B 95% / l 8 95% Cl long dimension N (lf) eekelii - sd N= sk 4 2 8 6 0 peekelii M. sd ( 21 20 19 18 17 16 15 14 13

2 ( 25 Part p 37 glaucum 2 ( N ) 16 0 glaucum crater dimension crater 21 20 19 18 17 16 15 14 13 midrib 3B Part anksii M. 7 E. sd

N= M. E. 22 16 22 25species 6 ( b 4A 7 3B Cl 17 25 ) species M. banksii M. peekelii M. peekelii Ensete sp. Ensete sp. M. anksii )

6 sd 24 ( b ) ) 22 (sd) (sd) (sd/fr) (lf) (sk) ) 95% 0 0 fr / 25 RATIO N= anksii sk 4 2 eekelii 8 6 0 sd sd N= (lf) 25 17 24 25 ( 7 M. N= 4B ( ( b 24 17 p 25 25peekelii 7 sd 25 Cl glaucum

5 25 glaucum ( eaf/stem 95% Cl crater dimension l M. banksiispeciesM. M.peekelii banksiiM.M. peekelii peekeliiE. M.glaucum peekeliiE. glaucumE. glaucum E. glaucum

species dimension crater E. species N= M. E. M. 7 6 5 4 3 2 1 0 M. 4A 17

(sd) (sd) 7 (sd/fr) (lf) (sk) Cl 95% 4 (sd) (sd) (sd/fr) (lf) (sk) N= 4 2 8 6 0

species 4A 4B 4A )

Figure 4. 95% confidence intervals24 for mean crater widths ) 95% Cl crater dimension 100 ) fr 95% Cl crater dimension95% anksii / 3 (lf)

of Musa spp. seed and Ensetesd glaucum leaf and fruit-skin species eekelii eed/fruit sd 4A ( b ( s glaucum peekelii p 25 sd

phytoliths based on full data set and with25 outliers deleted A ( E. M. crater dimension crater M. 2 B (sd = seed; fr = fruit; lf = leaf; skM. = skin). Musa acuminata 7 17 ssp. banksii is clearly differentiated at α = 0.05 whenCl outliers N= 4 2 8 6 0 ) 332 1 are removed. The Australimusa species M. peekelii is not fr 24 ) ) eaf/bract / 95% l

N= 332 100 25 37 species differentiated from Ensete glaucumanksii . 4A sd eekelii peekelii sd sd ( b 25 ( ( p Part leaf/bract seed/fruit leaf/stem midrib 25 9 8 7 6 5 4 3 2 1 N= crater dimension crater M. Part 95% Cl RATIO M. 3C M.

. 95% confidence intervals for mean crater widths , mean long dimensions and mean long dimension/ Cl Figure 3 A B 17 3C N= 4 2 mean crater width ratios C of all phytoliths examined.8 Note6 that the leaf/stem sample consists0 of the base of leaf and 24 ) 95% pseudostem samples and the seed/fruit samples consist of: A) the fruit pulp attached to seeds, and B) seeds.) Fruit pulp eekelii sd species anksii 4A sd (

does not contain phytoliths and therefore phytoliths examined in the analysis are derived from seeds only. Thep ratio ( b plot (C) shows that seed and seed/fruit phytoliths have significantly smaller craters compared to body25 length than leaf M. M. phytoliths and can be discriminated at α = 0.05 regardless of species derivation. 17 N= 4 2 www.ethnobotanyjournal.org/vol7/i1547-3465-07-247.pdf8 6 0 ) anksii sd species ( b 4A 25 M. N= 4 2 8 6 0 species 4A 254 Ethnobotany Research & Applications

seed morphotypes (see Figure 1) that were not includ- Analysis 2 ed in this analysis. However, the leaf and fruit skin phytolith morphotypes of Ensete have globular and polygo- This analysis was based on morphotypic analysis of the nal morphotypes similar to Musa seed morphotypes. An same set of leaf/bract volcaniform morphotypes used in analysis comparing mean body length and mean crater the first analysis. Seven major categories of attributes width of M. acuminata ssp. banksii and Musa peekelii (Table 2) were examined. Attributes within each catego- Lauterb. seed morphotypes, and E. glaucum leaf and ry were given a score of 1 if present and 0 if absent and fruit skin morphotypes showed that the width of craters statistically tested using principal components analysis. in M. acuminata ssp. banksii are significantly smaller Similar to the morphometric analysis, Ensete glaucum than both M. peekelii and Ensete at α = 0.05 (Figure 4). morphotypes were clearly differentiated by body texture M. peekelii and E. glaucum could not be differentiated and crater rim characteristics (Figure 6). Other taxa according to crater width. However, mean body length could not be clearly differentiated.

sp. of M. peekelii was significantly greater at α = 0.05 than ) M. acuminata ssp. banksii and Ensete (Figure 5). Analysis 3 ) sk ( sk ( Ensete 40 glaucum Additional morphometric analyses were undertaken to E. determine if leaf phytolith morphotypes could be further sp. differentiated. Mean body length and mean crater width (lf) ) of leaf/bract volcaniform morphotypes from different sk 30 (lf) ( taxa were compared (Figures 7 to 9). Crater width of Ensete glaucum glaucum 6 E. glaucum morphotypes was significantly smaller at α E. E.

7 = 0.05 than all other Eumusa and Australimusa banan- )

fr as with the exception of the wild Australimusa species, / 20 ) peekelii sd ) fr Musa maclayi F. Muell. ex Mikl.-Maclay and the wild Eu- 25 ( / (lf) sk longdimension musa species Musa schizocarpa Simmonds (Figure 7). ( M. peekelii glaucum sd 25 ( glaucum Cl Body length was significantly smaller than the two M. E. crater dimension crater M. E. 7 acuminata cultivars (listed as M. acuminata and Musa

10 Cl 22 ) 95% paradisiaca L.), M. maclayi and Fe`i (Figure 8). Crater ) eekelii sd ( 24 fr p )

) length and body length of morphotypes from other ba- 95% / (lf) eekelii sd sk peekelii

M. nanas were more similar and consequently these taxa sd ( ( 25 p glaucum ( 16 0 glaucum

crater dimension crater were found to be more difficult to differentiate (Figures 7 M. E.

N= M. 22 16 22 25E. 6 7

40 Cl 17 and 8). Crater width of the wild M. maclayi morphotypes ) species M. banksii M. peekelii M. peekelii Ensete sp. Ensete sp.

anksii were significantly different from the cultivated bananas, sd ) 24 ( b ) 22 (sd) ) (sd) (sd/fr) (lf) (sk) fr 95% ) M. acuminata and Fe`i and the other wild Australimusa / sk eekelii anksii sd M. (lf) sd 5 ( ( ( peekelii p b banana M. peekelii, but overlapped with the wild Eumu- sd glaucum 25 25 glaucum 95% Cl( long dimension

N= sa species (M. acuminata ssp. banksii and M. schizo- crater dimension crater 0 E. E. M. M. 30 M. 40 30 20 10 17 7

Cl carpa). Musa schizocarpa could only be differentiated species N= 4 2 8 6 0 5 from the three cultivars (M. acuminata, M. paradisiaca ) 24 ) ) fr Figure 5. 95% confidence intervals95% for mean long and Fe`i), and the wild M. acuminata ssp. banksii was (lf) / anksii sd species eekelii sd 4A ( b glaucum dimensions of Musa spp. seed and ( Ensete leaf and fruit- differentiated from only one of the Eumusa section cul- peekelii p 25 sd 25 20 (

skin phytoliths based on full data set (sd = seed; fr =E. fruit; tivars, M. acuminata, not M. paradisiaca. Body length M. crater dimension crater M. M. 7

lf = leaf; sk = skin). The Australimusa17 species, M. peekelii was a less helpful criterion for differentiating taxa than Cl N= 4 2 8 6 0 is differentiated from Musa acuminata ssp. banksii and) crater width. Interestingly, the only significant difference fr 24 ) ) /

Ensete glaucum. 95% at α = 0.05 was between Australimusa section banan- species anksii 4A sd eekelii peekelii sd sd ( b 10 25 ( ( p 25 crater dimension crater

Table 2. Categories and attributes used in analysis M. of volcaniform leaf phytoliths (attributes shown in Figure 6B are the M. M. Cl

abbreviated forms shown in (…) and are17 equivalent terms from Madella et al. 2005). N= 4 2 8 6 0 24

Categories Attributes ) 95% ) eekelii

0 sd species anksii ( 4A sd

Crater round [r] (orbicular), oval [o], squarep [sqt], irregular [irrt] ( b

N= 22 16 22 2525 6

Morphology tabular [t], blocky [b], spherical [sph]M. (globose), platy [pl] (planar) M.

species M. banksii M. peekelii M. peekelii17 Ensete sp. Ensete sp. Base shape square [squb], rectangular [rb], quadrilateral [qb], triangular [tb], boat [bb] (oblong), round [cb] N= 4 2 8 6 (sd) (sd) 0 (sd/fr) (lf) (sk)

(orbicular), other [ob] ) anksii

5 sd species ( b

Height short [fh] (h<1/3 length), medium4A [mh] (h=1/3 to 1/2 length), tall [th] (h≥1/2 length) 95% Cl long dimension25

Texture psilate [stx], rough [rtx] verrucate,M. granulate [grx], dimpled [dtx] N= 4 2 Rim 8 6 present [rp]/absent [ra], regular0 [regr]/irregular [irrr] Ornamentation absent [no], short [sho], medium [mo], long [lo], lobed [lbo] species 4A

www.ethnobotanyjournal.org/vol7/i1547-3465-07-247.pdf Lentfer - Tracing Domestication and Cultivation of Bananas from Phytoliths: 255 An update from Papua New Guinea as. Body lengths of M. maclayi and Fe`i bananas were the greatest variability and could not be differentiated significantly greater than M. peekelii. Finally the ratio from any other taxon. of mean body length to crater width differentiated M. maclayi from all other samples with the exception of Lentfer’s findings broadly concur with those of Ball et al. M. schizocarpa (Figure 9). Musa schizocarpa exhibited (2006) and Vrydaghs et al. (2009), which have been con- fined to the distinctive volcaniform leaf phytoliths from a different set of Musaceae accessions including M. acumi- A 3 nata, Musa balbisiana Colla and various cultivar groups. A combination of morphometric (base length and crater width) and morphotypic analyses (base shape, crater po- 2

8 sition and cone shape) (Ball et al. 2006:3) can help to discriminate between certain taxa. Notably, wild diploid M. 1 balbisiana (BB) volcaniform morphotypes were found to 6 be significantly larger than both wild and edible diploid M. acuminata (AA) morphotypes (Ball et al. 2006:7), but edible AA could not be differentiated from wild AA. Sub-

4 0 sequent studies analyzing AA, AAA, AAB and ABB have PC 2

found a very complex pattern of phytolithPC 2 variation. Con-

2 -1 tinuing analyses with additional samples are further inves-

tigating the variation in crater width,0.8 particularly the role of banksii alleles in its expression (Vrydaghs et al. 2009). 0 -2 NO PC 1 RA SPH Implications for future 0.6 research 2 - -3 FH The presence of seeded bananas in archaeobotanical as- -4 -2 0 2 4 6 8 semblages can be identified from seed phytolith morpho- 0.4 4 Rough texture Rough Irregular rim -

PC 1 types.B Distinctive morphotypes shown in Figure 1 are di- 1 2 3 3 2 1 0 - - -

Musa6A schizocarpa Simmonds IRRT

IRRR agnostic at the section level. Other globular and polygonal SQT PC 2 morphotypes can be differentiated from other plant parts Musa6A peekelii Lauterb. DTX 0.2 GRX by body length/crater width ratios. Additionally, prelimi- Musa paradisiaca L. SHO naryOB studies indicate that Eumusa bananas (M. acuminata RTX O

Musa maclayi F. Muell. ex Mikl.-Maclay QB ssp. banksii) can be differentiated from Australimusa ba- 0.0

nanasTB (M. peekelii) on the basis of crater width and body Musa fei F. Muell0.8. PC 1

CB IRRR RTXBB RB Musa acuminata Colla0.8 0.8 TH B RTX IRRR IRRR

RTX 0.2

MO Rough texture

Ensete sp. 0.6 - R IrregularRough STX rim texture Musa banksii F. Muell0.6.T 0.6 LO Rough texture R SHO Irregular rim Irregular rim 0.4 SQUB

T T TH 0.4 Figure 6. Biplot of principal R R SHO SHO - 0.4 REGR 0.4 MH components analysis of banana TH TH 0.2 CB QB RA leaf phytolith attribute data. MO SQUB RP GRX FH 0.2 RB CB QB CB RA0.6 The sample plot A shows that 0.2 QB - RA MO TB MOGRX FH FH Ensete glaucum is differentiated0.0 RP SQUB SQUB GRX MH T BBRB OB RBB from other bananas. The0.0 main TB DTX TB RP0.0MH LO RP BBO OB BB B OB B attributes separating it from MH SQTDTX NO 0.8

PC 2 DTX -0.2 - other bananas are the irregular LO LOO IRRT O 0.8 0.6 0.4 0.2 0.0 0.2 SQT0.4 0.6 0.8 SQTNO NO - - - - -0.2 6B rim and the rough texture -0.2 IRRT SPH 6Akey and to a lesser - extent,0.4 short 0.8 IRRT SPH SPH ornamentation (SHO).- 0.4 See the distribution of attributes in the -0.4 NO RA SPH -0.6 vector plot B. (see abbreviations 0.6 STX in Table 2) -0.6 -0.6REGR FH -0.8 REGR STX REGR STX

-0.8 -0.4 0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 Rough texture Rough Irregular rim -0.8 -0.8 B -0.8 -0.6 -0.8-0.4 -0.6-0.2PC-0.4 0.01 -0.20.2 0.00.4 0.20.6 0.40.8 0.6 0.8 IRRT

IRRR 6B

SQT PC 2 PC 1 PC 1

DTX 6B

0.2 6B PC 2 GRX PC 2 SHO OB RTX O

QB www.ethnobotanyjournal.org/vol7/i1547-3465-07-247.pdf 0.0 TB PC 1 CB BB RB TH 0.2 MO - R STX LO SQUB 0.4 - REGR MH RP 0.6 - T 0.8 - 0.8 0.6 0.4 0.2 0.0 0.2 0.4 0.6 0.8 - - - - 6B 256 Ethnobotany Research & Applications

length. Although further compara- 10 tive studies are needed to include a range of other species from either section, current results sug- gest the outlook is very promising 9 for tracking the complex history of Musaceae in the archaeobo- dimension crater tanical record. Phytoliths can be Cl used to identify natural distribu- 8 tions of Musa and Ensete, differ- 95% entiate wild populations from fully domesticated (seedless) popula- 7 tions and trace patterns of disper- dimension crater

sal. However, based on this prem- Cl ise, mixed populations of wild and

cultivated bananas (a common 95% 6 occurrence in Papua New Guin- ea; Lentfer 2003b; Jean Kenne- dy pers. comm.) and partially do- 5 mesticated populations prior to N= 58 25 25 25 76 25 50 50 the complete loss of seed – for species Figure 7. 95% confidence intervals for mean crater widths of leaf 7 phytoliths differentiating Ensete schizocarpa acuminata peekelii paradisiaca maclayi banksii glaucum from all other bananas fei glaucum except Musa maclayi and Musa M. M. M. M. M. M. E. schizocarpa. M. 7 labels Figure 8. 95% confidence intervals 22 schizocarpa acuminata peekelii for mean long dimensionsparadisiaca of leaf maclayi banksii fei glaucum phytoliths differentiating Ensete M. M. M. M. M. M. E. glaucum M. from four other bananas. 20

some diploid cultivars commonly 7 labels produce seed – cannot be differ- 18 entiated. In these circumstances it is only the presence of Musaceae longdimension species outside their natural rang- 16 Cl es that might imply human trans- 95%

mission and confirm evidence for longdimension cultivation. Cl 14 Where seed is absent from archaeobotanical assemblages, the 95% problem of identification becomes 12 more difficult and is reliant on differentiation of the distinctive volcaniform morphotypes. From the 10 results of several studies (Lentfer N= 58 25 25 25 76 25 50 50 et al. 2003b, Mbida et al. 2001) species it is well-established that Ensete species can be readily differenti- 8 ated from wild and domesticated diploid and triploid Musa bananas schizocarpa acuminata peekelii paradisiaca maclayi banksii by morphotypic and morphomet- fei glaucum ric means. However, differentiation M. M. M. M. M. M. E. M. between Musa species is more schizocarpa acuminata peekelii paradisiaca maclayi banksii fei

glaucum www.ethnobotanyjournal.org/vol7/i1547-3465-07-247.pdf M. M. M. M. M. M. E. M. 8titles 8titles Lentfer - Tracing Domestication and Cultivation of Bananas from Phytoliths: 257 An update from Papua New Guinea complex and would be reliant on 4.0 a large sample size for any given archaeobotanical assemblage. Since triploid banana phytoliths are generally larger than diploids 3.5 (Vrydaghs et al. 2009), there could be scope for differentiating

domesticated triploid populations (ld/td) ratio

from wild and cultivated diploid 3.0 Cl populations by simply measuring

crater widths of archaeological 95% ratio (ld/td) ratio assemblages. Therefore, there

Cl 2.5 is potential for tracking banana introductions and domestication. Nevertheless, results show that 95% Eumusa and Australimusa sec- 2.0 tion bananas cannot be differentiated at a general level and this is problematic in regions where 1.5 bananas from both sections oc- 25 25 cur, either wild or cultivated. Indi- N= 58 25 76 25 50 50 cations are, however, that some species species and/or cultivars within sections can be differentiated. 9 Most importantly, wild and domesticated Eumusa bananas, schizocarpa acuminata peekelii paradisiaca maclayi banksii fei

M. acuminata (AA) and M. bal- glaucum M. M. M. M. M. M. E. bisiana (BB), can be differenti- M. ated and there might be poten- schizocarpa acuminata peekelii paradisiaca maclayi banksii tial forfei tracking the introduction Figure 9. 95% confidence intervals for mean long dimension (ld)/crater width (td) glaucum of Musa acuminata bananas in ratios of leaf phytoliths differentiating Musa maclayi from all other bananas except M. M. M. M. M. M. E. M. mainland Southeast Asia west Musa schizocarpa. 9titles of the Philippines where M. balbisiana dominates native banana populations. A similar taken, describing more morphological features (Table 5).

potential for discrimination is indicated for9titles Australimusa; Digital images and measurements and morphometric de- M. maclayi and M. peekelii could not only be differenti- tails from 50 phytoliths per accession have been recorded ated from each other but also from the Australimusa do- and saved on a readily accessible database. This in itself mesticate Fe`i. Therefore, there may be good potential for is useful for identification of morphotypes during routine tracking Australimusa banana dispersals and patterns of analysis. Statistical analyses have yet to be completed. domestication for the near Oceania region, east of Papua Firstly, data will be lumped according to the same criteria New Guinea, where Australimusa bananas have dominat- as Ball et al. (2006) and examined using the same statis- ed wild and cultivated populations. tical procedure as Ball et al. (2006) and Vrydaghs et al. (2009) for direct comparison. Subsequently, the analysis The ‘New Guinea Banana Project’ will be re-run to incorporate the full set of morphotypic attributes and morphometries. Good potential for differentiating between banana phytoliths is indicated from morphometric and morphotyp- Conclusions ic analyses. Nevertheless, preliminary studies point to a wide variation of morphotypes and additional study of a Recent research (Denham et al. 2003, Lebot 1999, Per- larger sample incorporating additional species and culti- rier et al. 2009) shows that the New Guinea region has vars is required to determine the extent of this variation played a key role in the development of the domesticated and further explore the potential for a more definitive set of banana, and complex origins and multiple dispersals for criteria for differentiation. The ‘New Guinea Banana Proj- banana cultivars within the southeast Asian/Pacific region ect’ commenced in 2002 with collection of more than 100 are indicated (Kennedy 2008). Phytoliths hold the key to wild and cultivated bananas from mainland Papua New tracing the history of banana cultivation and domestica- Guinea, New Britain and New Ireland (Table 3). Volcani- tion in the archaeobotanical record, and morphometric form leaf phytoliths from 58 accessions were selected for and morphotypic analyses show good potential for differ- a more rigorous analysis (Table 4) than previously under- entiation of phytoliths. Seed phytoliths can be readily dif-

www.ethnobotanyjournal.org/vol7/i1547-3465-07-247.pdf 258 Ethnobotany Research & Applications Notes Seeds used for making beads. Cooked and also eaten ripe. Cooked and also eaten ripe. Cooked and also eaten ripe. Cooking banana. Amele Language Bakovi Bakovi Bakovi Bakovi Bakovi ? Towberne Sihir Rukumamb tamby Kunambo Gana sumdu Tagomor ? Lakem ? ? ? Sar Luba Vernacular Vernacular Name ? Tapupu Kikiyou Tukuru buro Tamane Maya II II II II II 3.6 3.6 3.6 3.6 3.6 I I I I I 5 5 5 5 5 o o o o o Longitude E 150 150 150 150 150 II II II II II 0.4 0.4 0.4 0.4 0.4 I I I I I 27 27 27 27 27 o o o o o 5 Latitude S 5 5 5 5 Lowlands Lowlands Lowlands Lowlands Highlands Highlands Lowlands Lowlands Lowlands Lowlands Lowlands Lowlands Lowlands Lowlands Lowlands Altitude (m) Lowlands Lowlands Lowlands Lowlands Lowlands Tamare Village Village Tamare East Sepik ENB Madang Western Western Highlands Chimbu Morobe ENB Madang (orig. Milne Bay) Madang Bougainville Central Province ? Madang ENB Cape Gloucester Location Tamare Village Village Tamare Village Tamare Village Tamare Village Tamare Ensete glaucum (Roxb.) Cheesman AA AAA AA AAB AAB ABB AA ABB AAA ABB AAB ABB AAT AAB ABB Type AA? ABB ABB? AA? WNB1 Nari 064 Nari OBN 14 Nari 164 Nari 206 Nari NBI 18 Nari NB9 11 Nari 142 Nari 171 Nari OBB 5 Nari 20 NBL Nari 168 Nari NBI 10 Nari 186 Nari 146 Nari 1047 Code WNB2 WNB3 WNB4 WNB5 Table 3 . List of banana accessions collected from Papua New Guinea in 2002 for the ‘New Banana Project’, with associated ethnobotanical observations. Table www.ethnobotanyjournal.org/vol7/i1547-3465-07-247.pdf Lentfer - Tracing Domestication and Cultivation of Bananas from Phytoliths: 259 An update from Papua New Guinea mumu . Notes Pseudostem & sheath used for making rope. Suckers sold to villagers by agricultural supply stores. Used as cash crop. Ornamental banana with red leaves. Eaten ripe. Cooked and also eaten ripe. Wild banana growing along margins of swamp. Collected from forest and grown in garden with other bananas. Necklaces made from seed. Sap red. Large seeds often present. Fruit cooked and also eaten ripe. Fruit also fed to pigs. Stems and leaves used for mumu . Fruit The making an alcoholic beverage. ripe fruit is placed in drum and after heating to release juices the liquid is fermented. It is distilled into a clean drum and put into small bottles. The process takes two weeks. Cooked and also eaten ripe. Fruit fed to pigs. Stems and leaves used for Fruit cooked. Male bud is also eaten. Eaten ripe and cooked. Eaten ripe. and torn easily Leaves banana. Cooking unsuitable for cooking and wrapping. Language Pisin Bakovi Bakovi English Pisin Kuanua Kuanua Kuanua Kuanua Kuanua Kuanua Kuanua

Vernacular Vernacular Name Tuain Hale vudi Karuke vudi ( Karuke hand) =crab’s or Mrs. banana sausage banana Wil banana vudu Vudu Yawa Kalapua Tukuru Kudukudu Kakatur Pok pok ll ll ll ll ll ll ll ll II II ll ll 29.4 27.2 02.4 02.4 27.1 27.1 59.0 48.6 l l l l l l l l 3.6 3.6 3.6 3.6 I I l l 5 5 5 5 58 27 10 10 10 10 08 14 o o o o o o o o o o o o Longitude E 150 150 150 150 149 150 152 152 152 152 152 152 ll ll ll ll ll ll ll ll II ll ll ll 0.4 0.4 0.4 0.4 12.9 25.2 47.0 47.0 58.9 26.3 58.9 38.0 I l l l l l l l l l l l 27 27 27 27 31 28 08 08 08 13 08 24 o o o o o o o o o o o o 4 Latitude S 5 5 5 5 5 5 4 4 4 4 4 Lowlands Altitude (m) Lowlands Lowlands Lowlands Lowlands Lowlands Lowlands Lowlands Lowlands Lowlands Lowlands Lowlands Tavui No. 2 Tavui Tamare Village Village Tamare Location Tamare Village Tamare Village Tamare Village Tamare Garu plantation Tabairikau No.2 Tavui No. 2 Tavui No. 2 Tavui Malaguna Gela

ABB Musa textilis Née Type ? AA? AA? Musa acuminata ssp. banksii (Ridl.) Kiew. E. glaucum ABB ABB ? ? ? ENB3 WNB6 Code WNB7 WNB8 WNB9 WNB10 WNB11 ENB1 ENB2 ENB4 ENB5 ENB6

www.ethnobotanyjournal.org/vol7/i1547-3465-07-247.pdf 260 Ethnobotany Research & Applications mumu . mumu . mumu . mumu . mumu , mumu only. mumu and eaten mumu and eaten raw. Notes Pseudostem and sheath used for making rope. Suckers sold by L.A.E.S. to agricultural supply stores. Cooking banana. Leaves easily torn and torn easily Leaves banana. Cooking unsuitable for cooking and wrapping. Cooking banana. Leaves used for mumu . and torn easily Leaves banana. Cooking unsuitable for cooking and wrapping. Eaten ripe. Leaves used for Eaten ripe. Leaves used for Eaten ripe. Leaves used for Eaten ripe and cooked. Eaten ripe and cooked. Leaves used for mumu . and torn easily Leaves banana. Cooking unsuitable for cooking and wrapping. Eaten ripe. Leaves used for Cooking banana. Eaten ripe and cooked. Cooking banana but also eaten ripe. Leaves strong and used for and wrapping. Male bud also eaten. Cooking banana - fruits boiled and roasted. Leaves weak - used for covering mumu only. Cooking banana - fruits boiled, fried and roasted. Leaves weak - used for covering Cooked in raw and boiled. Sweet. Cooked in Language Kuanua Kuanua Kuanua Kuanua Kuanua Kuanua Kuanua Kuanua Kuanua Kuanua Kuanua Kuanua Tigak Tigak Tigak Tigak Pisin , Vernacular Vernacular Name Makau Katur Pitu Makala tukuru buka Yawa Okaoko Auro = Vuro Vuro Gulum / Pitu Vuro Taupen Malam red Sukuru Papat (cf. Garamut in Kuanua) Kikiyou kina Wan Boa Wan pound Wan ll ll ll ll ll ll ll ll ll ll ll ll ll ll ll ll ll ll ll 51.6 49.1 49.1 3.1 18.3 3.1 54.6 54.6 54.6 54.1 54.1 l l l l l l l l l l l 31.2 41.0 41.0 48.2 48.2 48.2 48.2 52.1 l l l l l l l l 14 14 14 13 12 13 4 1 1 1 1 1 1 1 47 47 47 47 47 o o o o o o o o o o o o o o o o o o o Longitude E 152 152 152 152 152 152 152 152 152 152 152 152 152 152 150 150 150 150 150 ll ll ll ll ll ll ll ll ll ll ll ll ll ll ll ll ll ll ll 36.7 36.1 36.1 29.8 6.0 29.8 49.6 7.0 7.5 7.5 7.5 7.5 9.2 43.1 43.1 43.1 45.5 45.5 l l l l l l l l l l l l l l l l l l 40.5 l 24 24 24 22 23 22 5 17 20 20 20 20 20 20 33 33 33 33 33 o o o o o o o o o o o o o o o o o o o 4 4 4 4 2 2 Latitude S 4 4 2 4 4 4 4 4 4 4 4 2 2 Lowlands Lowlands Lowlands Lowlands Lowlands Lowlands Altitude (m) Lowlands Lowlands Lowlands Lowlands Lowlands Lowlands Lowlands Lowlands Lowlands Lowlands Lowlands Lowlands Lowlands Malakuna No. 4 Mission Takabua Station Raolo L.A.E.S. Kerevat Kulangit Village, Kavieng Kulangit Village, Kulangit Village, Kavieng Gela Location L.A.E.S. Kerevat Kulangit Village, Kulangit Village, Kavieng Gela Gela Mission Takabua Station Malmaluan College Training L.A.E.S. Kerevat L.A.E.S. Kerevat L.A.E.S. Kerevat L.A.E.S. Kerevat Kulangit Village, Kavieng Kulangit Village, Kavieng ABB BB? Feì AB? ABB? ? ? Type M. textilis ABB? cf. Tukuru ( Kokopo ) AAA AA ABB Feì AA Feì AAA AS? AA? AA ENB11 ENB12 ENB13 ENB17 NI4 NI5 ENB7 Code ENB20 NI1 ENB8 ENB9 ENB10 ENB14 ENB15 ENB16 ENB18 ENB19 NI2 NI3

www.ethnobotanyjournal.org/vol7/i1547-3465-07-247.pdf Lentfer - Tracing Domestication and Cultivation of Bananas from Phytoliths: 261 An update from Papua New Guinea mumu . mumu . Notes Eaten ripe and cooked. Yellow flesh Yellow Eaten ripe and cooked. with poorly formed seeds eaten Bunch urine yellow. Turns by ants. drooping. Possibly intermediate form. Leaves strong and used in Eaten ripe and cooked. Yellow flesh. Leaves not used. Eaten ripe and cooked. Yellow flesh. Leaves not used. Cooking banana - boiled and roasted. Leaves not continuous spiral around rachis. used. Fruit occurs Fruit in roasted or used. boiled. Leaves not used Leaves raw. eaten or cooked Fruit for wrapping. for used Leaves boiled. or roasted Fruit mumu and wrapping. Mostly roasted. Cooked and eaten raw - sweet. Sap used as a dye for malangan masks. Eaten ripe. Eaten ripe or cooked. Leaves used for mumu and wrapping. Bunch drooping but at an angle and almost horizontal. Eaten ripe or cooked. Leaves used for mumu and wrapping. Bunch drooping but at an angle and almost horizontal. Cooking banana. Flesh sulfur-yellow. Wild. Said not to be eaten but sweet bright yellow flesh surrounds seeds. Leaves used for wrapping and Language Tigak Tigak Tigak Tigak Tigak Tigak Tigak English/Tigak Malik Mandak Manui (Manus Island)/ Mandak Mandak Mandak Barok Vernacular Vernacular Name Suluklamuk Darip Palang Ngitingsakai Ulungan Bavaya Papat Wung White bavaya Utafan Siaman / German Manui/ Namnam Loskauk Loskauk Agigilai ll ll ll ll ll ll ll ll ll ll ll ll ll ll 54.1 54.1 52.1 52.1 56.4 56.1 41.4 41.4 57.1 39.0 22.9 48.7 30.2 46.9 l l l l l l l l l l l l l l 47 47 47 47 47 47 49 49 24 55 54 54 55 11 o o o o o o o o o o o o o o Longitude E 150 150 150 150 150 150 150 150 151 151 151 151 151 152 ll ll ll ll ll ll ll ll ll ll ll ll ll ll 45.5 45.5 41.7 41.7 42.5 48.4 25.6 25.6 20.7 28.8 58.3 55.4 52.7 l l l l l l l l l l l l l 45.4 l 33 33 33 33 33 33 35 35 2 16 19 17 16 29 o o o o o o o o o o o o o o 3 Latitude S 2 2 2 2 2 2 2 2 3 3 3 3 3 <1000m Altitude (m) Lowlands Lowlands Lowlands Lowlands Lowlands Lowlands Lowlands Lowlands Lowlands <1000m <1000m <1000m <1000m Karu Kulangit Village, Kulangit Village, Kavieng Location Kulangit Village, Kulangit Village, Kavieng Kulangit Village, Kavieng Kulangit Village, Kavieng Kulangit Village, Kavieng Miom/Caranas, Kavieng Miom/Caranas, Kavieng Namasalang Lelet Kaluan, Lelet Kaluan, Lelet Kabagong, Lelet Kulangit Village, Kulangit Village, Kavieng

Arg. Feì ? Type AAA AA AA AAB AA AAB Feì ? Feì Feì Feì Musa maclayi Muell. ex F. Mikl.-Maclay ssp. maclayi Arg. var. namatani ? NI6 Code NI8 NI9 NI10 NI11 NI12 N13 NI14 NI15 NI16 NI17 NI18 NI19 NI7

www.ethnobotanyjournal.org/vol7/i1547-3465-07-247.pdf 262 Ethnobotany Research & Applications mumu mumu . mumu . mumu . mumu . mumu . Notes Wild. Said not to be eaten but sweet bright yellow flesh surrounds seeds. Leaves used for wrapping and first at erect Bunch banana. robust Wild roasted. or cooked Fruit drooping. then Leaves used to cover first at erect Bunch banana. robust Wild roasted. or cooked Fruit drooping. then Leaves used to cover Wild tall banana growing 18m to height 16m at to 2000m in Fruit with yellow flesh and large seeds, creek gully. Cuscus, ripe. when yellow turning green parrots and flying fox eat fruit. Leaves used for making temporary and for covering shelters 18m to 16m to growing banana tall Wild height at 2000m altitude in creek gully. Fruit with yellow flesh and large seeds. Leaves used for shelters and for covering making temporary As above. Cooking banana, roasted and boiled - preferably roasted. Wild banana growing in rainforest with gingers, ferns, Rubiaceae shrubs, tree shade in rainforest in taro wild and ferns under canopy. Tall robust banana with when eaten Fruit wild. than rather Feral edible fruit. for used Leaves cooked. or raw and wrapping. Fruit also fed to pigs. Wild. Tall, robust banana with pseudostem. shiny Not fruiting at collection. time of Language Barok Sokiri Sokiri Jiwaka Jiwaka Jiwaka ? Mino Barok

= red) Vernacular Vernacular Name Agigilai Wawa Wawa donu Wuluk donu Wuluk donu Wuluk Rua mema ( rua = banana, mema ? (=wild Yesing banana) ll ll ll ll ll ll ll ll ll ll 21.2 l 3.6 l 3.6 3.6 3.6 3.6 3.6 3.6 3.6 3.6 l l l l l l l l 12 5 5 5 5 5 5 5 5 o o o o o o o o o Longitude E 152 150 150 150 150o5 150 150 150 150 150 ll ll ll ll ll ll ll ll ll ll 20.3 0.4 0.4 0.4 0.4 0.4 0.4 0.4 0.4 0.4 l l l l l l l l l l 30 27 27 27 27 27 27 27 27 27 o o o o o o o o o o 5 5 5 5 Latitude S 3 5 5 5 5 5 <1000m <1000m 2000m 2000m Altitude (m) <1000m 2000m 1630m 1177m Lowlands <1000m Rosirik Rosirik Kauil Tubang, Kauil Tubang, Karu Location Balga, Kuk Ranimbo, Hawain Baiyer Zoo Tubang, Kauil Tubang, Karu

Arg. Simmonds Musa peekelii Lauterb. ssp. peekelii M. peekelii ssp. peekelii Musa ingens M. ingens M. maclayi ? Fe`i ? Type Fe`i Musa schizocarpa AAT? Simmonds M. ingens M. maclayi ssp. maclayi namatani var. NI22 NI23 WH1 WH2 NI20 Code WH4 WH5 ES1 WH3 NI21 www.ethnobotanyjournal.org/vol7/i1547-3465-07-247.pdf Lentfer - Tracing Domestication and Cultivation of Bananas from Phytoliths: 263 An update from Papua New Guinea

singsing decoration as well as , wrapping food wrapping mumu , . This specimen growing specimen This . mumu . singsing mumu and dried leaves dried and Notes Fruit green turning yellow when in ripe. decoration as used Leaves Wild seeded banana sap. with yellowish Wild seeded banana with clear for used Leaves sap. and for roofs for bush houses. Cultivated banana Fruit large with 4-5cm diameter, 10-23cm red long with sap. pronounced ridging. Mostly cooked. bottle-necked apex and Wild seeded banana with clear Fruit sap. turns yellowish when landslides colonize often to Said eaten. ripe. Not with gingers and towards Highway Sepik along common grasses. Species with association in grows Often Maprik. schizocarpa M. on well-drained loam along roadside next to gardens and close to regrowth forest. One plant only. Cooking banana. Leaves wrapping and used for used Leaves cooked. or ripe eaten Fruit for used for making cigarettes. Clear sap. Fruit eaten ripe used for mumu . or cooked. Leaves Wild banana growing on sandy Fruit not eaten. bank. Sap clear. river regrowth shady in growing banana Wild forest with moss understory. celebrations after heating on make pliable. fire to Language Mino Mino Mino Mino Mino Pisin ?, Mino Mino, Madang language? Mino Mino Vernacular Vernacular Name (=wild Yesing banana) (=wild Yesing banana) (=wild Yesing banana) Taweyawa (=wild Yesing banana) Giant kalapua Salamua , pisang jari buaya (crocodile fingers) To`o Heifeli’ (=wild Yesing banana) (=wild Yesing banana) ll ll ll ll ll ll ll ll ll ll 29.4 27.2 02.4 02.4 27.1 27.1 59.0 48.6 51.6 49.1 l l l l l l l l l l 58 27 10 10 10 10 08 14 14 14 o o o o o o o o o o Longitude E 149 150 152 152 152 152 152 152 152 152 ll ll ll ll ll ll ll ll ll ll 12.9 25.2 47.0 47.0 58.9 26.3 58.9 38.0 36.7 36.1 l l l l l l l l l l 31 28 08 08 08 13 08 24 24 24 o o o o o o o o o o Latitude S 5 4 4 4 4 4 4 4 4 5 Altitude (m) Lowlands Lowlands Lowlands Lowlands <1000m <1000m <1000m <1000m <1000m Lowlands Location CCRI station, Hawain Ranimbo, Hawain Koikin Village, Wewak Koikin Village, Wewak Japaraga 1, Maprik Road Japaraga 1, Maprik Road Japaraga 1, Nagam River Japaraga 1, Nagam River Japaraga 1, Maprik Road CCRI station, Hawain Type AA? M. schizocarpa Fe`i M. acuminata ssp. banksii ABBT? AA M. acuminata x schizocarpa M. acuminata ssp. banksii AA M. schizocarpa Code ES2 ES4 ES5 ES6 ES7 ES9 ES10 ES11 ES8 ES3

www.ethnobotanyjournal.org/vol7/i1547-3465-07-247.pdf 264 Ethnobotany Research & Applications

M. Fe`i Musa peekelii mumu and wrapping . Common in well-drained ” in Pisin because urine because Pisin in pispis ” Fe`i bananas are referred to angustigemma . (NB. Site used for Notes Cultivated banana with red sap growing sap red with banana Cultivated on well-drained red/brown loam. Fruit large with pronounced cooked. Mostly flesh. sulfur-yellow ridging and bananas once common in the district often out, dying gradually be to said but settings. garden abandoned in surviving Notable “ yelo as Fruit eaten. is fruit the after yellow turns food other coloring for used and cooked sap Red flavours. enhancing and stuffs on stencils making for children by used clothing. Small cultivated although plantain as to referred banana banana. Cooking strictly speaking it isn’t. Leaves used for mumu . Cooking banana, boiled Leaves or used for roasted. food. drained poorly on growing banana Wild limestone soil in regrowth forest along roadside. Growing near near growing banana wild robust Tall ssp. making the film “Robinson Crusoe”). Fruit self-peels to appearance to give the apex. a Skin to used Leaves white. Flesh ripe. when green star-like make temporary bush shelters. schizocarpa locations on in regrowth coral/limestone forest. falling Bunch straight. soils Flowers eaten drooping, - mixed with salt. Young eaten opportunistically. pseudostem Flowers also pseudostem used to garnish taro. and Language Ari Amele Kuanua Em Em Vernacular Vernacular Name Wahin Manameg (red) Kekiau Maror Dor ll ll ll ll ll 49.1 3.1 18.3 3.1 l l l l 31.2 l 14 13 12 13 4 o o o o o Longitude E 152 152 152 152 152 ll ll ll ll ll 36.1 29.8 6.0 29.8 l l l l 40.5 l 24 22 23 22 5 o o o o o Latitude S 4 4 4 4 4 Altitude (m) cl Lowlands Lowlands Lowlands Lowlands Motonau/Midiba, Madang Location Stewart Research Station, Madang Stewart Research Station, Madang Baitata Road, Madang Baitata Road, Madang

Fe`i Type AA AA? M. schizocarpa M. peekelii ssp. angustigemma M1 Code M2 M3 M4 M5

www.ethnobotanyjournal.org/vol7/i1547-3465-07-247.pdf Lentfer - Tracing Domestication and Cultivation of Bananas from Phytoliths: 265 An update from Papua New Guinea M. a and and AS and banksii in angustigemma M. schizocarpa M. schizocarp M. schizocarpa and M. acuminata ssp. M. peekelii ssp. appearance. Male bud green. Notes hybrids. Wild banana growing in with association on coral/limestone soils at regrowth forest. edges of Wild banana in stand growing moderately on dense in regrowth well-drained forest along soil roadside. close In association with AS hybrids. and Wild banana growing on well-drained soil in regrowth forest along roadside. Growing amongst M. mixed acuminata , stand of Wild banana growing on well-drained soil in regrowth forest along roadside. Growing amongst mixed stand of Cultivated species the very wild much like Growing plantation. in Seeds necklaces. neglected used to Wild coconut make banana growing gardens. in Leaves sometimes used overgrown for covering food. Species also found at Gomila Village. purple-red milky dark, with banana Wild sap and erect bunch. Fruit with sulfur- sheltered cool, in Growing flesh. yellow gully in regrowth forest. Another small stand found growing nearby on rocky talis slope beside creek in old garden bordering regrowth forest. acuminata hybrids. Language Em Kein Kein Kein Kein Haigwai Haigwai Tawala

bihi = wild Vernacular Vernacular Name Tilja Mabul kulal ( mabul =wild: kulal =inedible, left to dry) Mabul jus ( jus = water snail because the snail has a pointed shell and the banana has a long pointed apex). Hirunag ( hiru =ginger; nag =son) Bagui Gudu gudu Lau moi ( lau =inedible; moi =banana) Bihibihiya ( bihibihi banana; = cultivated banana) ll ll ll ll ll ll ll ll 41.0 41.0 48.2 48.2 3.6 3.6 3.6 3.6 l l l l l l l l 1 1 1 1 5 5 5 5 o o o o o o o o Longitude E 152 152 152 152 150 150 150 150 ll ll ll ll ll ll ll ll 49.6 7.0 7.5 7.5 0.4 0.4 0.4 0.4 l l l l l l l l 17 20 20 20 27 27 27 27 o o o o o o o o 5 Latitude S 4 4 4 4 5 5 5 Lowlands Altitude (m) Lowlands Lowlands Lowlands Lowlands Lowlands Lowlands Lowlands Naura Village Baitata Village, Baitata Village, Madang Location Barum junction, Madang Barum junction, Madang Barum junction, Madang Naura Village Gomila Village, Ahioma Jobtou, Madang ? E. glaucum M. acuminata ssp. banksii Type M. schizocarpa M. acuminata ssp. banksii AS M. schizocarpa M. maclayi ssp. maclayi AA MB1 M6 Code M9 M7 M8 MB2 MB3 M10

www.ethnobotanyjournal.org/vol7/i1547-3465-07-247.pdf 266 Ethnobotany Research & Applications M. according maclayi Arg. but in other ssp. ailului ssp. M. maclayi M. Notes Specimen found on hillside in regrowth forest recently cleared for areas forest in gardening. growing be to noted Also inland from the coast. One of stand of dark, wild milky bananas purple-red with on talis sap slope in old garden growing alongside in found species of stands Small creek. dense regrowth forest on steep banks bordering creek. Wild banana with pink sap and erect bunch, tight but lax at proximal end. Fruit with sulfur-yellow flesh. Growing in gardens on raised sandy gravel deposits bordered by creek channels. Wild banana in small stand of bananas growing on well-drained roadside in tall grassland. soil along scattered small in growing banana Wild roadside along soil well-drained on stand in tall rachis grassland. the on bracts of retention feature Some specimens while others bract shed retention is a characteristic of bracts. Notably Said to be growing in bush catchment. Not collected. in creek Cultivated banana. Fruit eaten cooked or ripe. Cultivated banana. Fruit eaten cooked or ripe. maclayi closely more fit specimens the respects with key. Argent’s to Language Tawala Tawala Tawala Tawala Daiamoni Daiamoni Daiamoni

bihi bihi bihi bihi = wild = wild = wild = wild Vernacular Vernacular Name Bihibihiya ( bihibihi banana; = cultivated banana) Bihibihiya ( bihibihi banana; = cultivated banana) Bihibihiya ( bihibihi banana; = cultivated banana) Bihibihiya ( bihibihi banana; = cultivated banana) Lau moi ( lau =false; moi =banana) Kokoluyu Masi masi ll ll ll ll ll ll ll ll 29.4 27.2 02.4 l l l 3.6 3.6 3.6 3.6 3.6 l l l l l 5 5 5 5 5 58 27 10 o o o o o o o o Longitude E 150 150 150 150 150 149 150 152 ll ll ll ll ll ll ll ll 0.4 0.4 0.4 0.4 0.4 12.9 25.2 47.0 l l l l l l l l 27 27 27 27 27 31 28 08 o o o o o o o o Latitude S 5 5 5 5 5 5 4 5 Altitude (m) Lowlands Lowlands Lowlands Lowlands Lowlands Lowlands Lowlands Lowlands Gomila Village, Gomila Village, Ahioma Location Tue Village Tue Gomila Village, Gomila Village, Ahioma Daduwe area, Northeast Coast Road Daduwe area, Northeast Coast Road Village Tue Village Tue Kuiaro Village, Kuiaro Village, China Strait ? ? ? ? ? ? M. maclayi ssp. maclayi Type M. maclayi ssp. maclayi M. maclayi ssp. maclayi M. maclayi ssp. maclayi M. maclayi ssp. maclayi AA AS? AA? M. maclayi ssp. maclayi MB4 Code MB0 MB5 MB6 MB7 MB8 MB9 MB0

www.ethnobotanyjournal.org/vol7/i1547-3465-07-247.pdf Lentfer - Tracing Domestication and Cultivation of Bananas from Phytoliths: 267 An update from Papua New Guinea Table 4. Accessions selected for the ‘New Guinea Banana Project’ analyses. Section Number Accession code analyzed Wild species Eumusa Musa acuminata ssp. banksii 5 ES11, ES6, M7, M6, ENB10 Musa schizocarpa 5 ES3, WH5, MB2, M9, ES4 Musa acuminata ssp. banksii x schizocarpa 2 M8, ES10 Australimusa Musa peekelii 2 M5, NI22 Musa maclayi 5 NI21, MB7, MB4, MB3, MB5 Musa textilis 2 ENB20, WNB6 Ingentimusa Musa ingens 3 WH1, WH2, WH3 N/A Ensete glaucum 4 WNB1, WNB11, MB1, MB1/2 Cultivars* Eumusa AA 5 NI12, ENB13, Nari064, Nari164, MB8 AAA 5 NI8, NariOBB5, NariNB420, NariOBN14, ENB18 BB? 1 ENB12 AAB 5 NI13, Nari206, Nari146, NI11, NB1 ABB 5 Nari1047, NariNBG11, NariNBI10, Nari171, NariNBL20 AB 1 ENB17 Australimusa TT(Fe`i) 5 WH4, ES5, ENB13, M1, ENB14 Eumusa x AAT 2 Nari186, ES1? Australimusa ABBT? 1 ES7 *Genome labels for diploid, triploid and polyploid cultivars: A = acuminata, B = balbisiana, T = Australimusa.

Table 5. Expanded list of categories and attributes used in analysis of volcaniform leaf phytoliths for the ‘New Guinea Banana Project’. Category Attribute Base, 3D shape tabular, blocky, spherical( globose), platy (planar) Base, 2D shape round, oval, square, rectangle, quadrilateral, boat (oblong), irregular Body length - Body width - Body height short (h<1/3 length), medium (h=1/3 to ½ length), tall (h≥1/2 length) Crater length - Crater width - Crater shape (dorsal view) round (orbicular), oval, quadrilateral, irregular Rim present, absent Rim shape regular, irregular, skirt (crenate) Sides straight, convex, concave, straight/concave, concave/convex, straight/convex Body texture psilate, granulate, verrucate, nodulose, tuberculate, psilate/verrucate, psilate/granulate, granulate/verrucate Base ornamentation absent, short (<1.25µm), medium (1.25-2.5µm), long (>2.5µm), long/tuberculate/ dendritic, short/medium, medium/long

www.ethnobotanyjournal.org/vol7/i1547-3465-07-247.pdf 268 Ethnobotany Research & Applications ferentiated from phytoliths generated by other plant parts, lecting Expeditions, 1988–1989. International Network for and the absence of seed phytoliths in an archaeobotani- the Improvement of Banana and Plantain, Montpellier. cal assemblage can signal presence of seedless/domesticated bananas. However, parthenocarpy and sterility, Ball, T., L. Vrydaghs, I. van den Houwe, J. Manwaring & E. processes inherent in banana domestication, evolved De Langhe. 2006. Differentiating banana phytoliths: Wild over several generations, and populations of seeded cul- and edible Musa acuminata and Musa balbisiana. Journal tivated diploid bananas are still common in Papua New of Archaeological Science 33:1228-1236. Guinea. Therefore, it can be difficult to determine the sta- tus of bananas (wild or cultivated) in the archaeobotanical Burkill. I.H. 1935. A Dictionary of the Economic Products record unless there is unequivocal evidence for bananas of the Malay Peninsula. Crown Agents for the Colonies, either being outside their natural range, associated with London. archaeological and pedogenic features indicative of cultivation, and/or associated with other known cultigens. In Carreel, F, D. González de Léon, P. Lagoda, C. Lanaud, the absence of diagnostic seed phytoliths identification is C. Jenny, J.-P. Horry & H. Tézenas du Montcel. 2002. As- reliant on volcaniform leaf phytoliths. Studies show that certaining maternal and paternal lineage within Musa by some species and cultivars can be discriminated accord- chloroplast and mitochondrial DNA RFLP analyses. Ge- ing to crater size and body length and by the presence nome 45:679-692. of certain rare morphotypes. Nevertheless, there is large morphotype variation within and between species; conse- Crowther, A. 2009. Investigating Lapita Subsistence and quently, identification is currently reliant on large sample Pottery Use Through Microscopic Residues on Ceram- sizes rarely encountered in fossil assemblages. There- ics: Methodological issues, feasibility and potential. Ph.D. fore, more work is needed to clarify the extent of variability Thesis, School of Social Science, Queensland University. across the geographic range and within different habitats. The ‘New Guinea Banana Project’ commenced in 2002 Daniells, J., C. Jenny, D. Karamura & K. Tomekpe. 2001. with the field collection of samples and resultant analysis Musalogue: A Catalogue of Musa Germplasm. Diversity in of 58 additional accessions. The outcome of these analy- the Genus Musa. International Network for the Improve- ses should help to resolve many of the outstanding issues ment of Banana and Plantain, Montpellier. regarding the differentiation of volcaniform leaf phytoliths in the archaeobotanical record, particularly in the Pacific/ De Langhe, E. & P. de Maret. 1999. Tracking the banana: Papua New Guinea region. its significance in early agriculture. Pp. 377-396 in The Prehistory of Food: Appetites for change. Edited by C. Acknowledgements Gosden & J. Hather. Routledge, London.

I am most grateful to the Pacific Science Foundation De Langhe, E., L. Vrydaghs, P. de Maret, X. Perrier & T. and the University of Queensland for providing the fund- Denham. 2009. Why bananas matter: An introduction to ing for this project. My sincere thanks to Clair Harris for the history of banana domestication. Ethnobotany Re- her assistance. I also thank Jeff Daniells, Bill Boyd and search and Applications 7:165-177. Robin Torrence who assisted with the field collection and the PNG government officials and staff (NARI, DAL, DPI, Denham, T.P. & M. Donohue. 2009. Pre-Austronesian dis- CCEA and CCRI) in particular Valentine Kambori, Rosa persal of banana cultivars west from New Guinea: linguis- Kambuou and Jeff Wiles for their support and kind assis- tic relics from eastern Indonesia. Archaeology in Oceania tance with getting the project up and running, and Janet 44:18-28. Paofa, Martin Gunther, David Minemba, James Maima, Moses Woruba, Will Akus, Joseph Wasem, Marcel Toko- Denham, T.P., S.G. Haberle, C.J. Lentfer, R. Fullagar, J. rewaga, Kauwe Murley, Luke Nama, Louis Kurika, March Field, M. Therin, N. Porch & B. Winsborough. 2003. Ori- Tovue, Joy Penias, Scola Singit, Nick Eki and Manau Peni gins of agriculture at Kuk Swamp in the Highlands of New for their kind assistance. I also thank Nick Lyons for his Guinea. Science 301:189-193. kind hospitality and all the people who showed us around their gardens and allowed us to take samples. Donohue, M. & T.P. Denham. 2009. Banana (Musa spp.) domestication in the Asia-Pacific region: Linguistic and ar- Literature Cited chaeobotanical perspectives. Ethnobotany Research and Applications 7:293-332. Argent, G.C.G. 1976. The wild bananas of New Guinea. Notes from the Royal Botanic Garden, Edinburgh 35:77- Horrocks, M., S. Bedford & M. Spriggs. In press. A short 114. note on banana (Musa) phytoliths in Lapita, immediately post-Lapita and modern period archaeological deposits Arnaud, E. & J.P. Horry. 1997.Editors of Musalogue: A from Vanuatu. Journal of Archaeological Science. Catalogue of Musa Germplasm. Papua New Guinea Col-

www.ethnobotanyjournal.org/vol7/i1547-3465-07-247.pdf Lentfer - Tracing Domestication and Cultivation of Bananas from Phytoliths: 269 An update from Papua New Guinea Horrocks, M., S. Bulmer & R.O. Gardner 2007. Plant mi- nice P. Bishop Museum Bulletin 190. Bernice P. Bishop crofossils in prehistoric archaeological deposits from Yuku Museum, Honolulu. rock shelter, Western Highlands, Papua New Guinea. Journal of Archaeological Science 35:290-301. Madella, M., A. Alexandre & T. Ball. 2005. International code for phytolith nomenclature 1.0. Annals of Botany Jarret, R.L., N. Gawel, A. Whittemore & S. Sharrock. 1992. 96:253-260. RFLP-based phylogeny of Musa species in Papua New Guinea. Theoretical and Applied Genetics 84:579-584. Mbida, C.M., H. Doutrelepont, L. Vrydaghs, R.L. Swen- nen, R.J. Swennen, H. Beeckman, E. De Langhe & P. de Kennedy, J. 2008. Pacific bananas: Complex origins, mul- Maret. 2001. First archaeological evidence of banana cul- tiple dispersals? Asian Perspectives 47:75-94. tivation in central Africa during the third millennium before present. Vegetation History and Archaeobotany 10:1-6. Kennedy, J. 2009. Bananas and people in the homeland of genus Musa: Not just pretty fruit. Ethnobotany Research Mbida, C.M., H. Doutrelepont, L. Vrydaghs, R.L. Swen- and Applications 7:179-197. nen, R.J. Swennen, H. Beeckman, E. De Langhe, & P. de Maret. 2004. Yes, there were bananas in Cameroon more Lebot, V. 1999. Biomolecular evidence for plant domes- than 2000 years ago. InfoMusa 13:40-42. tication in Sahul. Genetic Resources and Crop Evolution 46:619-628. Neumann, K. & E. Hildebrand 2009. Early bananas in Af- rica: The state of the art. Ethnobotany Research and Ap- Lejju, B.J., P. Robertshaw & D. Taylor. 2006. Africa’s earli- plications 7:353-362. est bananas? Journal of Archaeological Science 33:102- 113. Perrier, X., F. Bakry, F. Carreel, Ch. Jenny, J.-P. Horry, V. Lebot & I. Hippolyte. 2009. Combining biological ap- proaches to shed light on evolution of edible bananas. Lentfer, C.J. 2003a. Tracing Antiquity of Banana Cultiva- Ethnobotany Research and Applications 7:199-216. tion in Papua New Guinea: Report on collection of modern reference material from Papua New Guinea in 2002. Un- Pollefeys, P., S. Sharrock & E. Arnaud. 2004. Preliminary published report prepared for the Pacific Biological Foun- Analysis of the Literature on the Distribution of Wild Musa dation, Sydney, NSW, Australia. Species using MGIS and DIVA-GIS. International Network for the Improvement of Banana and Plantain, Montpelli- Lentfer, C.J. 2003b. Plants, People and Landscapes in er, and International Plant Genetic Resources Institute, Prehistoric Papua New Guinea. A compendium of phyto- Rome. lith (and starch) analyses. Ph.D. Thesis, Southern Cross University, Lismore, N.S.W., Australia. Sharrock, S. 1990. Collecting Musa in Papua New Guin- ea. Pp. 140-157 in Identification of Genetic Diversity in the Lentfer, C.J. 2009. Going bananas in Papua New Guin- Genus Musa: Proceedings of an International Workshop ea: A preliminary study of starch granule morphotypes in held at Los Baños, Philippines, 5–10 September 1988. Musaceae fruit. Ethnobotany Research and Applications Edited by R.L. Jarret. International Network for the Im- 7:217-238. provement of Banana and Plantain, Montpellier.

Lentfer, C.J. & R.C. Green. 2004. Phytoliths and the evi- Simmonds, N.W. 1959. Bananas. Longmans, Green and dence for banana cultivation at the Lapita Reber-Rakival Co, London. site on Watom Island, Papua New Guinea. Pp. 75-88 in A Pacific Odyssey: Archaeology and anthropology in the Simmonds, N.W. 1962. The Evolution of the Bananas. Western Pacific. Papers in Honour of Jim Specht. Records Longmans, Green and Co, London. of the Australian Museum 29. Edited by V. Attenbrow & R. Fullagar. Australian Museum, Sydney. Spriggs, M. 1996. What is Southeast Asian about Lapita? Pp. 324-348 in Prehistoric Mongoloid Dispersals. Edited Lentfer, C.J., C. Pavlides, R. Fullagar & J. Specht. 2008. by T. Akazawa & E. Szathmary. Oxford University Press, Subsistence and Environmental History in Central New Oxford. Britain, Papua New Guinea: Combining phytoliths, macro- fossils and use-wear/residue studies. Paper presentation Valmayor, R.V. 2001. Classification and characterization at the Australian Archaeological Conference, Noosa, Qld. of Musa exotica, M. alinsanaya and M. acuminata ssp. er- December 2008. rans. InfoMusa 10:35-39.

MacDaniels, L.H. 1947. A Study of the Fe`i Banana and its Vrydaghs, L., T. Ball, H. Volkaeart, I. van den Houwe, J. Distribution with Reference to Polynesian Migrations. Ber- Manwaring & E. De Langhe. 2009. Differentiating the vol-

www.ethnobotanyjournal.org/vol7/i1547-3465-07-247.pdf 270 Ethnobotany Research & Applications caniform phytoliths of bananas: Musa acuminata. Ethno- Wong, C., R. Kiew, J.P. Loh, L.H. Gan, O. Set, S.K. Lee, botany Research and Applications 7:239-246. S. Lum & Y.Y. Gan. 2001. Genetic diversity of the wild banana Musa acuminata Colla in Malaysia as evidenced by Wilson, S.M. 1985. Phytolith analysis at Kuk, an early ag- AFLP. Annals of Botany 88:1017-1025. ricultural site in Papua New Guinea. Archaeology in Oce- ania 20:90-97.