INFORMATION TO USERS

This material was producad from a microfilm copy of tha original document. While the most advanced technological means to photograph and reproduce this document have been used, the quality is heavily dependant upon tha quality of the original submitted.

Tha following explanation of techniques is provided to help you understand markings or patterns which may appear on this reproduction.

1. The sign or "target" for pages apparently lacking from tha document photographed is "Missing Page(s)". If it was possible to obtain the missing paga(s) or section, they are spliced into tha film along with adjacent pages. This may have necessitated cutting thru an image and duplicating adjacent pages to insure you complete continuity.

2. Whan an image on tha film is obliterated with a large round black mark, it is an indication that the photographer suspected that the copy may have moved during exposure and thus causa a blurted image. You will* find good image o f tha page in the adjacent frame.

3. When a map, drawing or chart, etc., was part of the malarial being photographed die photographer followed a definite method in "sectioning" the material. It is customary to begin photoing at the upper left hand corner of a large sheet and to continue photoing from left to right in equal sections w ith a small overlap. If necessary, sectioning Is continued again — beginning below tha first row and continuing on until complete.

4. Tha majority of users indicate that tha textual content is of greatest value, however, a somewhat higher quality reproduction could be made from "photographs" if essential to the understanding of tha dissertation. Silver prints o f "photographs" may be ordered at additional charge by writing the Department, giving the catalog number, tide, author and specific pages you wish reproduced.

5. PLEASE NOTE: Some pages may have indistinct print. Filmed at received.

University Microfilms International 300 North 2««b Road Ann Arbor, Michigan 48106 USA S t John's Road, Tyler's Green High Wycombe, Bucks, England HP10 BHR I I 77-31,992 STOCKEY, Ruth Anne, 1950- MORPHOLOGY AND REPRODUCTIVE BIOLOGY OF CERRO CUADRADO . The Ohio State University, Ph.D., 1977

University Microfilms International, Ann Arbor. Michi0an woe MORPHOLOGY AND REPRODUCTIVE BIOLOGY OF CERRO CUADRADO

FOSSIL CONIFERS

«

DISSERTATION

Presented in Partial Fulfillment of the Requirements for

the Degree of Doctor of Philosophy in the Graduate

School of The Ohio State University

r

By

Ruth Anne Stockey, B.Sc., M.Sc. ******

The Ohio State University

1977

Reading Committee:

Dr. Thomas N. Taylor Dr. Gary L. Floyd Dr. Barbara A. Schaal Dr. Valaymghat Raghavan

Adviser Department of Bota: This publication is dedicated to the memory of the late Dr. Hans Tralau who provided much assistance and many of the specimens during this study, and whose untimely passing constitutes a great loss to the study of paleo­ botany.

ii ACKNOWLEDGEMENTS

The author wishes to thank the following people for the acquisition and loan of specimens and the use of lab­ oratory facilities: Dr. Eugene S. Richardson, Field

Museum of Natural History, Chicago; Mr. Cedric Shute,

British Museum of Natural History, London: Dr. Britta

Lundblad, Naturhistoriska Riksmuseet, Stockholm; Dr.

James M. Schopf and Dr. Stig Bergstrom, United States

Geological Survey Coal Geology Laboratory and Orton Mu­ seum, The Ohio State University; Dr. John W. Hall, Uni­ versity of Minnesota; Dr. Francis M. Hueber, United States

National Museum, Smithsonian Institution; Dr. Elizabeth

McClintock, California Academy of Sciences, San Francisco;

Mr. Frederick C. Boutin, Huntington Botanical Gardens,

San Marino, CA; Dr. Enrique C. Clos, Jardin Agrobotanico de Santa Catalina, Buenos Aires; Dr. Garth Nikles,

Forestry Department, Brisbane; Dr. Richard Gould and Mr.

Ron Larson, Imbil Forestry Station, Queensland; M. Fran­ cois Guinaudeau, C. T. F. T. Parc Forestier, Noumea, New

Caledonia. Special appreciation is expressed to Dr.

Thomas N. Taylor for his guidance and support during this study. Supported in part by National Science Foundation grants BMS 75-14221 and BMS 74-21105.

iii VITA

January 3, 1950 Born - Harvey, Illinois

1972-1974. . . Teaching Assistant, Ohio University, Athens, Ohio

1974 ...... M.Sc., Ohio University, Athens, Ohio

1974-1977 Research and Teaching Assis­ tant The Ohio State Uni­ versity Columbus, Ohio

PUBLICATIONS

Stockey, R. A. 1974. Seeds and Of mirabilis. Amer. J. Bot. 61s 19-20 (Abstract). ______. 1975a. Pararaucaria patagonica Wieland from the Cerro Cuadrado petrified forest () : Seeds and embryos. Akad. Nauk SSSR (XII International Botanical Congress Abstracts) p. 237.

______. 1975b. Seeds and embryos of Araucaria mirabilis Amer. J. Bot. 62:856-868.

______. 1975c. Morphology and anatomy of Pararaucaria patagonicas Seeds and embryos.' Bbt. Soc. Amer. Abstracts. Allen Press, Lawrence, Kansas p. 25.

. 1975d. Immature ovulate cones of Araucaria mirabilis. Bot. Soc. Amer, Abstracts'! Allen Press, Lawrence, Kansas, p. 25.

______, and T.N. Taylor. 1976. Supposed silicified from the Cerro Cuadrado petrified forest (Jurassic) . Bo.t. Soc. Amer. Abstracts. Allen Press, Lawrence, Kansas, p. 32.

Taylor, T. N. and R. A. Stockey. 1975. The morphological nature of Microspermopteris. Bot. Soc. Amer.

iv Abstracts. Allen Press, Lawrence, Kansas, p. 25- 26.

______, and . 1976. Studies of Paleozoic seed ferns: Anatomy and morphology of Microspermopteris aphyllum. Amer. J. Bot. 63:1302-l3l0,

Stockey, R. A. 1977. Reproductive biology’of the Cerro Cuadrado (Jurassic) fossil conifers: Pararaucaria patagonica Wieland. Amer. J. Bot. 65:734-746,

FIELDS OF STUDY

Major Field:

Phylogeny and reproductive biology of the Araucaria- ceae and fossil, conifer groups. Professor Thomas N. Taylor.

Carboniferous coal ball permineralized : anatomy, morphology, ontogeny, and phylogenetic trends. Professor Thomas N. Taylor.

v LIST OF FIGURES

Page

Figures 1-7. Pararaucaria patagonica...... 103

Figures 8-13. Seed and cone-scale features. . . . . 105

Figures 14-20. Pararaucaria patagonica embryos. . . 107

Figures 21-27. Seedlings ...... 109

Figures 28-33. Seedlings ...... Ill

Figures 34-37. Seedlings ...... 113

Figures 38-44. Seedlings ...... 115

Figures 45-53. Seedlings ...... 117

Figures 54-57. Seedlings ...... 119

Figures 58-68. Araucaria mirabilis...... 121

Figures 69-76. Araucaria mirabilis...... 123

Figures 77-82. Araucaria mirabilis...... 125

Figures 83-90. Araucaria mirabilis...... 127

Figures 91-96. Araucaria mirabilis...... 129

Figures 97-101. Araucaria mirabilis...... 131 TABLE OP CONTENTS

Page

ACKNOWLEDGEMENTS...... iii

VITA...... iv

LIST OF F I G U R E S ...... vi

INTRODUCTION...... 1

MATERIALS AND METHODS ...... i . . . 5

Directory of Specimens...... 5 Techniques...... ; ...... 9

STRATIGRAPHY ...... 14

PARARAUCARIA PATAGONICA WIELAND ...... 16

Introduction...... 16 Emended diagnosis ...... 17 Description...... 19 General features ...... 19 Cone axis...... 20 Cone-scale complex ...... 21 Seeds ...... 23 Embryos...... 27 Discussion...... 29

SEEDLINGS ...... 41

Introduction...... 41 Fossil seedlings...... 42 Living araucarian seedlings ...... 44 Other fossil seedlings...... 49 Discussion...... 52

ARAUCARIA MIRABILIS (SPEGAZZINI) WINDHAUSEN EMEND. C A L D E R ...... 59

Introduction...... 59 Emended diagnosis . '...... 63 Description . . . 64

vii Page

Cone structure...... 64 Cone a x i s ...... ‘ 66 canals...... 67 Isolated cone axes...... 69 Cone-scale complex...... 70 and seeds...... 71 Integuments. . . . ‘...... 72 vascularization...... 73 N u c e l l u s ...... 74 Abortive ovules...... 76 Megaspore membrane ...... 77 Megagametophy te...... 77 Embryos . . . k ...... 78 Discussion ...... 81

GENERAL DISCUSSION ...... 96

APPENDIX ...... 103

LIST OF REFERENCES...... 133

viii INTRODUCTION

The Cerro Cuadrado Petrified Forest of Patagonia has yielded some of the best fossil conifer remains known.

Silicified cones, wood, twigs with leaf scars, and sup­ posed "seedlings” from the province of Santa Cruz, Argen­ tina have been scattered to various universities and mu­ seums around the world. The original material comes from a forty square mile area in southern Patagonia near the bases of three Volcanic peaks, the Cerro Cuadrado, Cerro

Alto, and Cerro Madre e Hija. Since the 1930's specimens have been sold as curios and kept in private collections because of their beauty and excellent preservation

(Wieland, 1935).

The petrified forest was initially described by

Windhausen (1924) who first collected silicified wood, seedlings and cones from the deposits in 1919 and 1922. His collection was passed on to Professor

Walter Gothan of who later described some of the specimens (1925, 1950). Unfortunately, ma ny of these originally described specimens have disapp eared. At about the same time a large collection was made by members of the Field Museum of Natural History's Patagonian expedition (Riggs, 1926). Cones and branches were re­ portedly found in place associated with many upright trunks. Riggs (1926) believed the were preserved on the site where they had grown. This Chicago collection was later reviewed by Wieland (1929, 1935) and incorporated in the Field Museum collections. A few later. Dr.

Franz Mansfeld collected and purchased many specimens while on a fossil vertebrate expedition in 1936 (Gordon,

1936; Calder, 1953). The massive collection of the Bri­ tish Museum (Natural History) London contains over 650 specimens from the Cerro Cuadrado region, most of which were purchased in 1936 from Mansfeld. In the Stockholm

Naturhistoriska Riksmuseet there are over 130 cones and hundreds of small twigs, many bearing leaf scars. Un­ countable additional specimens have been sold to European and American universities.

The initial studies of Cerro Cuadrado conifers were done by Spegazzini (1924), Gothan (1925), and Wieland

(1929, 1935). Wieland's monograph in particular included a description of the remains found in the large Field

4 Museum collection as well as a discussion of the process of silicification. He described other well known petrified * forests throughout the world and included a discussion of ’ extant araucarian forests in general. Later studies in­ cluded those by Gothan (1950), Calder (1953), and Menendez 3

(1960). Calder (1953) has provided the most critical study to date of these from the collections of the Bri­

tish Museum. She reviewed the literature on the Cerro

Cuadrado forest and presented revised diagnoses on twigs and* branches (Araucarites santaecrucis Calder) and ovulate cones Pararaucaria patagonica Wieland, Araucaria' mirabilis . .

(Speg.) Windhausen. Menendez (1960) described three pollen cone specimens associated with the Cerro Cuadrado remains.

Unfortunately, the poor preservation of these cones made it difficult for him to assign them to any living of conifers. Consequently, the binomial Masculostrobus altoensis Menendez is used for such specimens.

The excellent silica permineralization of these coni­

fer remains has been noted by Darrow (1936) who published a short paper on one cone in which she illustrated the general features of the of Araucaria mirabilis.

Darrow, like other previous Cerro Cuadrado workers, used polished slabs for her investigation with few thin sections;

the cellular organization of the cone, seed, and embryo

tissues was not discussed.

Through the use of extensive thin sectioning, and

other techniques, the present investigation considers

the cellular organization of the ovulate cones of

Araucaria mirabilis and Pararaucaria patagonica, seeds

as well as embryo tissues. An attempt has been made not

I 4 only to provide a detailed anatomical analysis of the cones, but to study these fossil conifers ontogenetically through use of a large number of specimens. The supposed

"" structures are also considered. It is the intent of this investigation to elucidate the development and reproductive biology of the Cerro Cuadrado fossil con­ ifers: to detail reproductive mechanisms as well as para­ meters of life cycle, ontogeny, and to closely compare them with extant genera in the and other conifer families. MATERIALS AND METHODS

This study utilizes fossil specimens borrowed from

five separate collections: Field Museum of Natural His­

tory (Chicago), British Museum (Natural History) London,

Naturhistoriska Riksmuseet (Stockholm), Orton Museum of

Geology (The Ohio State University, Columbus), and the

. University of Minnesota (Minneapolis).

Directory of Specimens

Field Museum of Natural History (Chicago)

Araucaria mirabilis

P13813 P13852 P13895 P13925 P13818 P13854 P13896 P13926 P13819 P13855 P13897 P13927 P13820 P13857 P13898 P13929 . P13827 P13858 P13901 P13930 P13828I P13859 P13903 P13933 P13828II P13860 P13904 P13935 P13828III P13862 P13906 P13937 P13829 P13864 P13907 P13939 P13830 P13865 P13908 P1393- P13836 P13867 P13909I P13972 P13837 P13868 P13909II P13973 P13838 P13880 P13914 P13981 P13840 P13881 P13915 (cone axis) P13841 P13883 P13917 P13846 P13885 P13918 P13847 P13890 P13920 P13848 • P13891 P13921 P13851 P13892 P13924 4 misc. pieces and 1 unnumbered cone (P15000)

5 Pararaucaria patagonica

P139— I P13963 P139--II P13967 P139— III P13969 P13920 P13970I P13941 P13970II P13948 P13976I P13951 P13976II P13952 P13978 P13953 P13979 P13954 P13982 P13957 P13985 P13958 P13986 P13959 P13988 P13961

"Seedlings'*

P14614 A P14614 B

British Museum (Natural History) London

Araucaria mirabilis

BM1052 BM1062 BM1053 BM1063 BM1054 BM1064 BM1055 BM1065 BM1056 BM1066 BM1057 BM1067 BM1058 BM1068 BM1059 BM1069 BM1060 V.31414 BM1061

Pararaucaria patagonica

BM1001 BM1011 BM1021 BM1031 BM1002 BM1012 BM1022 BM1032 BM1003 BM1013 BM1023 BM1033 BM1004 BM1014 BM1024 BM1034 BM1005 BM1015 BM1025 BM1035 BM1006 BM1016 BM1026 BM1036 BM1007 BM1017 BM1027 BM1037 BM1008 BM1018 BM1028 BM1038 BM1009 BM1019 BM1029 BM1039 BM1010 BM1020 BM1030 BM1040 Pararaucaria patagonica (cont.)

BM1041 BM1042 BM1043 BM1044 BM1045 BM1046 BM1047 BM1048 BM1049 BM1050 BM1051

"Seedlings"

BM1070 BM1077 BM1071 BM1078 BM1072 BM1079 BM1073 BM1080 BM1074 BM1081 BM1075 BM1082 BM107 6

Naturhistoriska Riksmuseet (Stockholm)

Araucaria mirabilis

SR2002 SR2012 SR2003 SR2013 SR2004 SR2014 SR2005 SR2015 SR2006 SR2016 SR2007 SR2017 SR2008 SR2018 SR2009 SR2019 SR2010 SR2020 SR2011 SR2021

Pararaucaria patagonica

SR2049 SR2056 SR2050 SR2057 SR2051 3R2058 SR2052 SR2059 SR2053 SR2060 SR2054 SR2061 SR2055 SR2062 Orton Museum

Araucaria mirabilis

031205 031221A 031206 031221B 031207 031221C 031208 031221D 031209 031221E 031210 031221F 031211 031221G 031212 031221H 031213 0312211 031214 031221J 031215 031221L 031221M

Pararaucaria patagonica

031216 031217 031218 031219 031220

University of Minnesota

Araucaria mirabilis

UM3002 UM3011 UM3003 UM3012 UM3004 UM3013 UM3005 UM3014 UM3006 UM3015 UM3007 UM3016 UM3008 UM3017 UM3009 UM3018 UM3010 UM3019

Pararaucaria patagonica

UM3001 Techniques

Specimens were selected for sectioning on examina­ tion of previous cuts made by earlier workers. Cones are completely silicified by a-quartz as determined by x-ray diffraction (Stockey, 1975a). Darrow (1936) points out the tendency of cones from these localities to have dif­ ferent colors of quartz replacing different seed tissues.

Embryos and seeds appear to be replaced by chalcedony, while the lumen once occupied by the nucellus is filled with transparent quartz.

Some of the specimens of Araucaria mirabilis and some wood pieces from the British Museum and Stockholm collec­ tions are embedded within the original volcanic ash matrix.

The ash seems relatively porous and contains some small pieces of charred remains. Some of the cones are embedded in nodular concretions, which upon cutting, re­ veal some of the best preserved specimens. Twigs, wood and cones often show evidence of the mode of burial by exhibiting scorched areas, pitting of the external sur­

face, exposure of seeds and distortions due to crushing

prior to silicification.

Slides were made with'2 x 3 in. double weight glass.

TRA-Bond 2114 Water White Transparent Epoxy Adhesive (TRA-CON, Inc., Medford, Massachusetts) was used for mounting specimens- to frosted glass slides. This epoxy resin has a refractive index (1.565) close to that of quartz thus giving excellent optical qualities. Few specimens showed little contrast with most tissues ap­ pearing clear-buff in color. Thin sections of- these specimens were stained in a 5% aqueous Malachite Green solution (Bartholomew# Matten, and Wheeler, 1970) for one and one-half minutes. Details of tracheal pitting and some integumentary and megagametophyte cells of Araucaria mirabilis showed a greater optical contrast utilizing this technique. Aqueous Bismark Brown stain was also tried on several sections, however, results were generally un­ satisfactory,

A number of living conifer seed wings were studied for comparison with those of Pararaucaria. These included seed wings of several genera of the Pinaceae,

Taxodiaceae and Araucariaceae.

Pinaceae

Abies amabilis Douglas ex Forbes

A. balsamea (L.) Miller

A. fraseri (Pursh) Poiret

A. grandis Lindley

A. lasiocarpa (Hooker) Nuttall

A. procera Rehder Larix laricina (Du Roi) K. Koch

Picea engelmannii (Parry) Engelmann

P. sitchenais (Bongard) Carriere

Plnus jeffreyi A. Murray

P. lambertiana Douglas

P. lumholtzii Robinson and Fernald

P. palustris Miller

P. pinea L.

. P^_ ponderoaa Douglas

P. radiata D. Don

P. rigida Miller

Pseudotsuga menziesii (Mirbel) Franco

Cupressaceae

Calocedrus decurrens (Torrey) Florin

Chamaecyparis lawsoniana (A. Murray) Parlatore

C. nootkatensis (D. Don) Spach

Taxodiaceae

Sequoia sempervirens (D. Don) Endlicher

Sequoiadendron giganteum (Lindley) Buchholz 12 Araucariaceae

Agathis australis (D. Don) Salisbury

A. dammara (A.B. Lambert) L.C. Richard

A. lanceolata (Pancher) Warb.

A. moorei (Lindley) Masters

Seed wings were softened in 5% NaOH at 37° c for two weeks, washed briefly in distilled water, then cleared

several days in chloral hydrate at 37° C at a concentra­

tion of 50g chloral hydrate to.30 ml water (Owens and Smith,

1965). Wings were washed overnight, stained with safranin,

* and dehydrated to . 100% EtOH and xylene, for subsequent mounting on slides with HSR (Harleco Synthetic Resin).

Dried seed wings of Abies grandis Lindley were coated with O approximately 200 A of gold and examined with a Hitachi

S-500 scanning electron microscope at 20 KV.

In an attempt to accurately determine the morphologi­

cal affinities of a wide variety of structures that have

been regarded as "seedlings" associated with the Cerro

Cuadrado fossil conifer remains, seedling structure of

from three sections of the Araucaria was

studied by means of paraffin (Johansen, 1940) and glycol methacrylate embedded sections (Feder and O'Brien, 1968,

Robison and Miller, 1975). Viable seeds from the sections

Eutacta, (A. columnaris (G. Forst) Hook.— 20 specimens; 13

A. cunninqhamll Ait. ex Lamb.— 50 specimens; A^ heterophyl- la (Salisb.) Franco — 100 specimens;' A. luxurians (Brongn. et Gris.) de Laubenfels — 59 specimens; A/ montana Brongn. et Gris.— 272 specimens; A/ rulei F. Muell.-- 240 speci­ mens; A;_ subulata Vieillard — 25 specimens ), Columbea

(A. angustifolia (Bertol.) O.Ktze ~ 15 specimens), and

Bunya (A. bidwillii Hook. — 35 specimens) , were soaked in tap water for 24 hrs. and planted in sterilized soil ap­ proximately 1/2 in. below the surface. * seeds were planted approximately 3 in. below the soil and covered with Sphagnum to retain moisture. Seedlings were . harvested at various stages of development during a 12 month period. Specimens were washed, photographed, and fixed in FPA prior to dehydration and preparation for sectioning. First- seedlings of Araucaria bidwillii,; obtained from the Imbil Forestry Station in southern

Queensland, were embedded and sectioned for anatomical study. STRATIGRAPHY

The age of the petrified forest is still not accur­ ately determined. An age determination of Middle was first suggested by Windhausen {1924, 1931) and sub­ sequently accepted by Gothan (1925) and Wieland (1935).

Gordon (1936) speaking for Mansfeld at the Geological

Society of London meetings expressed the opinion that the forest was Tertiary in age based solely on the modern appearance of the araucarian cones found there. Florin

(1940, 1944) believed the volcanics were probably Creta­ ceous in age. Frenguelli (1933) followed by Darrow

(1936) and Fossa-Mancini (1941) suggested the deposits were Eocene. The stratigraphy of a series of compression floras from a number of outcrops in the area between

Rio Deseado and Rio Chico in Santa Cruz suggests a Middle to age (Feruglio, 1949, 1951). This-inter­ pretation is followed by Calder (1953) , Menendez (I960), and Stockey (1975a). The major unconformity that separ­ ate the volcanic sediments containing cones from the

Upper rocks above, as well as the presence of the branchiopod genus Cyzicus Audouin (=Estheria Ruppell) in the sediments below, lend additional support to

14 Peruglio's conclusions because Cyzicus draper! (Jones)

Audouin is generally accepted as a Late Triassic index fossil (Feruglio, 1949). PARARAUCARIA PATAGONICA WIELAND

Introduction

The silicified ovuliferous cone Pararaucaria patagonica was first described by Wieland (1929) in a ♦ preliminary treatment of the Field Museum (Chicago) collection. This cone is the smaller of the two ovulate cones known from the petrified forest and at first was thought to represent two distinct species, P^ patagonica and P_;_ elongata (Wieland, 1929). No figure accompanied

Wieland's preliminary report and therefore these names were not validly published. In 1935, however, Wieland himself demonstrated that the range of characters in specimens of Pararaucaria intergraded and did not warrant the recognition of two species, since the variation within the cones was no more than that occurring in extant Pinus cones. This initial valid description with illustrations was based on 20 cone specimens of the Field Museum collec­ tion through the use of polished surfaces. Wieland be­ lieved the cone represented a transition form between the

Araucariaceae and the Pinaceae based on seed and cone- scale features and cone vascularization. He placed it in

16 the Cheirolepidacea, a family of conifers

M ■ * * erected by Hirmer and Horhammer . (1934) and subsequently emended by Jung (1968).

Calder emended the generic diagnosis of Pararaucaria * to fit the range of variation seen in the British Museum collection, and employed thin sectioning in addition to studies of polished surfaces. Xylem structure in-the cone axis was compared to extant conifer families. In addi­ tion, the presence of many in the embryo, and the winged .nature of the seed were reported. Calder placed Pararaucaria in the Taxodiaceae because of the great diversity within the family (sensu Pilger, 1926) stating that it resembled no single living genus in the combination of characters present.

Pararaucaria patagonica Wieland 1935. p. 21. pis.

2-5.

SYNONOMY

1929. Pararaucaria patagonica Wieland p. 60. nomen invalidum (no published figure).

1929. Pararaucaria elongata Wieland p. 60. nomen invalidum (no published figure).

* Emended diagnosis

Pararaucaria patagonica Wieland — Ovulate cones, conical to cylindrical in shape, varying in length from 2.3-5.1 cm and from 1.3-2.6 cm in diameter, borne on peduncles bearing 1 8 spirally arranged, imbricate, broadly lanceolate leaves with acute tips and longitudinal striations on the abaxial

surfaces. Cone-scales composed of woody, flattened

subtending axillary ovuliferous scales, usually 38 in

number. and ovuliferous scale conspicuous, arranged

in a 3/8 phyllotaxy.. Bracts approximately 11 mm long, 12

mm wide, 1.5.-.3.0 mm thick, free from fertile scale for

greater part of length; ovuliferous scale approximately

12 mm long, 15 mm wide, 2.5 - 4.5 mm thick, showing long­

itudinal ridges on distal abaxial surface; bearing 1 or

rarely 2 seeds, flattened, winged, heart-shaped, laterally

inserted; the seed separating from scale at maturity,

having abscission zone at chalaza. Seed approximately 7 mm

long, 7 mm wide, 2.5 mm thick, with wing about 2 mm wide to tapering toward base. Integument multi-layered, inner

endotesta of mucilage or resin filled parenchyma, middle

sclerotesta 3-4 layered, isodiametric — rounded sclereids

overlain by 1-2 rows closely packed radially elongate

sclereids. Wing and upper scale surface with anastomosing

glandular hairs with intercellular spaces. Embryo approx­

imately 4 mm long, 6-8 cotyledons, little megagametophyte

present at maturity.

Cone axis 3 mm in diameter, with narrow pith, and

thick endarch xylem cylinder. Primary xylem succession,

scalariform to reticulate to biseriate circular-bordered

pits. Secondary xylem usually with uniseriate 19 circular-bordered pits, contiguous and flattened, pit apertures round or ovate; medullary rays with cross field pitting of cupressoid form. Resin ducts and cells absent.

Circular bract trace single from axis stele, ovuliferous scale trace formed by fusion of 2 inverted traces, abaxi- ally concave, accompanied by large sclerenchyma strand that forks into.2 triangular bundles. Bract and ovuliferous scale traces single near axes, branch distally; abundant transfusion tissue linking bundles tangentially and per­ sisting in tips. Seed vascular supply single.

DESCRIPTION

General features — Cones vary in shape from conical to cylindrical, sometimes ovoid, and are approxi­ mately' 5.1 - 2.3 cm long and 2.6 - 1.3 cm in diam (Fig. 1).

Numerous helically arranged cone-scale complexes are arranged in a 3/8 phyllotaxy around a slender cone axis.

Each cone-scale complex is composed of a large woody ovuliferous scale which is subtended by a smaller woody, flattened bract (Fig. 1, 2). Externally cone-scales i resemble those of certain members of the Taxodiaceae, e.g., Sciadopitys Siebald and Zuccarini and Seguoiadendron

Buchholz; however, the bract and ovuliferous scale of

Pararaucaria are not fused but free for almost their entire length (Fig. 3,6). In less weathered specimens (Fig. 2), the outer surface of the ovuliferous scale appears striated. No f laminar tip, umbo, or extension of the ovuliferous scale into distinct lobes has been observed. Some cones show evidence of scorched - areas or pitting of the ovuliferous scales and distortion due to crushing, suggesting evidence of burial by volcanic ash. In a few weathered cones seeds with visible internal tissues are exposed. Only four specimens of those examined show pieces of cone peduncle,. and all of these appear to have poorly preserved leaves of the type described by Calder (1953) surrounding the axis. She figured one British Museum specimen with spiral­ ly arranged, broad, lanceolate leaves with striated sur­ faces that show a sharp transition to fertile scales.

No cuticle remains are present.

Cone axis — The pith of the cone axis is small

(0.5-0.8mm) and often not well preserved (Fig. 5).

Cells appear small, thin-walled, isodiametric (45 urn diam) .

Some apparent cellular contents and small spherical

(5-6 um diam) structures are common inside the best pre­ served cones. Larger cells about 65 um in diam with dark brown contents are scattered throughout the pith. In longitudinal sections small, scattered, squared-rectangular sclereids up to 80 um in diam can be seen. This type of . . 21 pith structure is similar to that in Taiwania cryptomer- ioides Hayata (Doyle and Doyle, 1948) .

Maturation of the primary xylem is endarch. There are helical wall thickenings on the innermost tracheids and these are succeeded by reticulate pitting. Some tracheids show biseriate, circular-bordered pits between the helical bands.

The wide zone of secondary xylem is relatively un­ dissected and composed of small diametered (20-30 um) tracheids (Pig. 7). Pitting on the radial walls is uni- seriate and circular-bordered. Vascular rays range from

1-5 cells high; no ray tracheids were observed. Resin canals and growth rings are absent. Crushed cells, often appearing in radial rows to the outside of the xylem, may constitute a zone of cambium, secondary phloem, and cortex

(Fig. 5). In the outer cortical region are groups of thick-walled sclerenchyma fibers that follow the vascular­ ization of the cone-scale complex out into the ovuliferous scale (Fig. 5, 11).

Cone-scale complex — Traces to the bract and its associated ovuliferous scale are separate at their origins

from the axis stele (Fig. 4). Thick woody bracts, each with one large vascular trace, are characteristic of

Pararaucaria. The bract is free from the associated ovuliferous scale for most of its length and fused near 22

the cone axis. Each bract trace is a circular to oval

in cross section for about half the length of the bract.

Farther out in the bract and ovuliferous scale are num­ erous isodiametric tracheids with circular-bordered pits

linking the vascular strands, making up what Calder (1953) refers to as a bulky transfusion tissue. The outer tis­

sues of the bract and -scale contain large cells 120-180 um in diam that are filled with brown .amorphous contents

(Fig. 3). Both surfaces of the bract appear to be smooth.

The abaxially concave ovuliferous scale trace is formed by the fusion of two separate strands that bend approximately 180°, and result in the protoxylem directed

toward the bract trace on the abaxial surface of the bundle.

As the ovuliferous scale trace departs from the axis stele

it is accompanied on the adaxial side by two massive

triangular shaped sclerenchyma strands that follow the

trace out into the ovuliferous scale (Fig. 5, 11). These

sclerenchyma strands originate from the dissection of one mass of sclerenchyma in the cortex of the cone axis. They remain in a triangular configuration about half-way out

into the scale, spread out laterally, and eventually dis­

appear distally. These sclerenchyma strands contain ex­

tremely thick-walled fibers 45 um in diam.

The upper and lower surfaces of the ovuliferous scale

are covered with what appear to have been glandular

trichomes. They are short cells that occasionally branch. 23

Similar hairs have been reported by Chowdhury (1961) on

the surfaces of scales of Cedrus, and it is believed that

they function in cone closure after pollination by the production of a resinous substance that hardens to seal * the cone closed. No actual resin canals have been found

in the bracts or ovuliferous scales of Pararaucaria.

» Seeds — Tissues of the seeds are well preserved in

several specimens. One flattened, winged seed is embedded

in the upper surface of each ovuliferous scale (Fig. 6).

Seeds are roughly heart-shaped in outline and about 6 mm

long.

One seed per cone-scale complex is characteristic

of araucarians, hence, the name Pararaucaria (Wieland,

1935). Seed integuments are usually represented by

wedges or fan-shaped bundles of cells in transverse section

(Fig. 10, 17-20). The integument of the seed is multi-

layered (Fig. 8) as is common in the Taxodiaceae and the

Cupressaceae at later stages of development (Konar and

Banerjee, 1963). The innennost cells compose an inner

fleshy layer of large cells filled with a black amorphous

material which may have been a mucilage or resinous sub­

stance in the living seed (Fig. 8).

The stony layer of the seed is composed of many

layers of cells (Fig. 8). The inner layer consists of

small rounded sclereids in transverse section that measure 24

35 um in diam. They are overlain by a layer of tightly packed radially elongate sclereids (65 um long) which is usually one cell thick. Some seeds show 2 layers of radi­ ally elongate cells in this region. Konar and Banerjee

(1963) have reported a similar type of seed integument in the extant Cupressus funebris in which sclerotestal cells are rounded sclereids at early stages of seed development and become radially elongate-with time. The formation of an additional layer of elongate sclereids in

Cupressus funebris marks the seed integument at maturity.

A surface view of the seed of Pararaucaria shows that the sclereids are oriented in longitudinal rows over most of the seed surface and not in isolated groups or bundles.

If preserved as a compression fossil the seed of Pararau­ caria would probably appear to have a striated surface similar to that reported for seeds of Elatides williamsoni by Harris (1943) from the Jurassic of Yorkshire, and

Elatocladus ramonensis'by Lorch (1967) from the Jurassic of Israel.

External to the stony layer are the possible remains of a fleshy layer of cells which appear crushed in most cases (Fig. 8). Cells in this region*when they are pre­ served are 20 um in diam and relatively thin-walled.

Seed wings are small, and measure a maximum of 2 mm

in length. They are composed of anastomosing rows of 25 glandular hairs with patches of thin-walled cells and many intercellular spaces (Fig. 9). These hairs are continuous with those on the upper surface of the ovuli-. ferous scale. A similar type of seed wing to that of

Pararaucaria is found only in Abies grandis (Fig. 12) whose surface reveals a,series of anastomosing cells, some that end in rounded tips such as. those seen in Pararaucaria seed wings and ovuliferous scale surfaces. Most of the taxa examined have wings composed of 1-3 layers of elon­ gate straight-walled cells, or cells with slightly wavy margins with no intercellular spaces or anastomoses.

Seeds of Araucaria are typically wingless.

The nucellus is preserved in most seeds as a thin layer of .crushed cells that attach only at'the seed chalaza. This tissue appears wavy in outline in most seeds.

The megaspore membrance is thin, appears crushed when present; and remains of megagametophyte cells are often present as a few rows of slightly compressed cells

(Fig. 14, 15, 18, 19). Cells are 60-90 um in diam, con­ tain granular appearing cell contents, and often spherical cellular inclusions up to 5 um in diam. Cells do not

show any evidence of alveoli as in .TaxujB megagametophyte

types (Chamberlain, 1935), but do show a regular arrange­ ment of isodiametric thin-walled calls. This type of 26 cell structure would indicate the presence of a Pinus type of megagametophyte (Chamberlain, 1935). None of the seeds thus far examined contains solid megagametophyte tissue without an embryo preserved inside; therefore, the inner-

t ' most cells of this tissue are unknown. No archegonia have been observed in any of the seeds.

The elongate; slender micropylar tube is oriented toward the cone axis and tangent to it. One vascular strand enters the seed from the distal portion of the ovuliferous scale (Fig. 6, 13). There is a thinning of the stony layer in the region of attachment and a pad of thin-walled parenchyma cells just outside the integuments which Calder (1953) referred to as a possible abscission zone (Fig. 13, arrow). Outside this zone is a group of small diameter tracheids which usually have a lobed, rec­ tangular configuration in cone longitudinal sections

(Fig. 13).

Wieland (1935) described a cone of Pararaucaria patagonica (P13939) that contained 2 seeds per ovuliferous scale, and in this study another specimen (P13979) has also yielded 2 seeds per scale (Fig. 10). Externally these cones appear identical with some of the larger

Pararaucaria specimens. Seed integuments, cone vascular­ ization and other features of these cones are indisting­ uishable from other cones of this taxon, suggesting that 27 such variation occurred within the genus in low frequency.

Mitra (1927) and Wilde and Eantes (195.5) point out varia­ tions in seed number in the genus Araucaria which has one seed per scale in most cases but in rare instances may have two to three seeds per scale. In cones of

Pararaucaria with two seeds per scale, there is no massive interseminal ridge as occurs in Pseudoaraucaria (Fliche,

1896; Alvin, 1953, 1957a, b).

♦

Embryos — The majority of seeds are sterile; how­ ever, where embryos are present, they are exceptionally well-preserved. The embryos of Pararaucaria patagonica are well-developed (Fig. 14-20). The shoot apex, coty­ ledons, hypocotyl, root meristern, and calyptroperiblem are all present indicating that embryos may have been in the dormant state when fossilized (Fig. 14). Some cellular remains of the megagametophyte occur around the embryo proper with crushed nucellar tissue surrounding them.

There is no provascular tissue apparent in this section, but it does appear to be present in transverse sections of some cotyledons. Cellular detail of the embryo tissue is well preserved. Delicate cell walls as well as cellu­ lar contents and small spherical structures (5-6 um diam) like those in pith cells are present (Fig. 16), The protoderm especially, covering hypocotyl and cotyledons, shows excellent preservation of cell walls and the small 2 8 spherical structures (Fig. 16). These structures might

have represented.a number of things in the living state:

protein bodies, starch grains, oil globules, other

stored food material, condensed cytoplasm, or most proba­ bly nuclear material. Their size range and position with­

in the cells would seem to favor the latter view. Addi­

tional support for this interpretation is their similarity

and presence in a number of different tissues; megagame­

tophyte, embryo, and pith parenchyma cells.

The shoot apex is about 200 um in diameter at the point of attachment of the cotyledons and has one promi­

nent surface layer of cells also with cellular contents.

The hypocotyl extends for a length of 1.8 mm below the

shoot apex and cotyledonary node. The generative meristern

of the root can be distinguished in the fossil by the change in orientation of the cells from hypocotyl to

calyptroperiblem (Fig. 14). The calyptroperiblem or "root

cap" (Buchholz and Old, 1933) extends beneath the root

meristem for a length of 0.9 mm (Fig. 14, 20). Also

apparent in the fossil embryos, is the column or columellar

region that in living conifer embryos adds cells to the

"root cap" by periclinal divisions (Allen, 1947).

Figures 17-20 represent a series of transverse

sections taken through the embryo of Pararaucaria. -Figure

20 represents a transverse section through the tip of the

calyptroperiblem in which the cells can be seen converging 29 toward a central point, the column or columellar region.

Figure 19 is a section through the embryo hypocotyl, and at a higher level (Fig. 18) through the cotyledonary node.

The indication of a three-parted symmetry and the presence of three distinct lobes or cotyledonary primordia are seen in Fig. 18, a section from an embryo which had 6 cotyledons. Figure 17 shows an embryo with 8 cotyledons, probably the result of the dissection of each of four cotyledonary primordia. There are 6-8 cotyledons per embryo, eight being the common number. This number corresponds to embryos in the Pinaceae (2-15) and the Taxodiaceae (2-9) (Butts and Buchhols, 1940; Bierhorst,

1971). Araucarians commonly have 2-4 cotyledons per embryo

(Burlingame, 1915). No resin canals have been seen in embryo transverse sections as have been described in the hypocotyl of Araucaria mirabilis from this locality

(Stockey, 1975a).

DISCUSSION

With the advent of more recent investigations of living and fossil conifers, and in particular cone vas­ cularization and resin canal distribution (Miller, 1969,

1970, 1972, 1973,*1974a, b; 1976; Miller and Robison, 1975), it is possible to compare more thoroughly this fossil genus with those of both extant and extinct conifer families. Historically/ Pararaucaria patagonica has been placed in

a number of conifer families by different authors.

Wieland (1935) believed it to be intermediate between the

Pinaceae and the Araucariaceae on the basis of seed num­

ber, large expanded cone-scale complexes, and cone vas­

cularization. Calder (1953) placed it in the Taxodiaceae

based on a number of characters, including the large con­

spicuous bract and ovuliferous scale, lateral seed attach­ ment, winged nature of.the seed, and cotyledon number.

Archangelsky (1968) placed Pararaucaria in the.Cheirolepi-

daceae (Hirmer and Horhammer, 1934), a family of Mesozoic

conifers containing about 6 genera. Wieland (1935) sug­

gested the Cheirolepidaceae for Pararaucaria due to the

peculiar combination of characters of the cone; Hirmerella

muensteri (Schenk) Jung comb. nov,. («=Cheirolepidium muen-

steri (Schimper) Takhtajan) and Tomaxellia Archangelsky

contain 2 seeds per scale. Pararaucaria has 1 or 2 seeds

per scale, the usual number being one. Members of the

Cheirolepidaceae, however, are characterized by having

prominently lobed cone-scale complexes. Since most of

the specimens of Pararaucaria studied thus far have under­

gone some weathering, the tips of the ovuliferous scales

or cuticle are not present for comparison. In cones where

there is a large portion of the ovuliferous scale present,

the scale appears striated and unlobed ('Fig. 2). Compressed

cones of Tomaxellia have the same general dimensions as 31

those of Pararaucaria, but shed their scales and retained

their, seeds at maturity like extant araucarians (Archangel­

sky, 1968).

Calder's (1953) table of positive correlation with

some families and genera of Coniferales compared Pararau-

4 , caria to the extant Araucariaceae, Pinaceae, Taxodiaceae,

and Cheirolepidaceae. Many characters are difficult to correlate to any degree because of the types of information

available due .to differences in preservation type among

the fossils and the current state of knowledge of living

conifers. Members of the Cheirolepidaceae, for example,

are preserved as compressions with cuticular remains and are difficult to compare in many case? with the petrified remains of Pararaucaria.

Externally, the cone resembles genera of the

Taxodiaceae, in particular Sciadopitys and Sequoiadendron.

The degree of fusion of the bract and ovuliferous scale is more pronounced in these genera, which show prominent

bracts (Hirmer, 1936). The Pinaceae, on the other hand,

are usually characterized by an inconspicuous bract (Hir­ mer, 1936; Miller, 1976). The ovuliferous scale of Para­

raucaria is the dominant structure within the cone-scale

complex. This feature, unlike the situation seen in

araucarians, is similar to that in the Pinaceae and the

Taxodiaceae. Eames (1913) points out that bract or scale 32 dominance is a character that varies in the Taxodiaceae, in particular in the genus Athrotaxis Don. Within this genus, A^ cupressoides Don shows an ovuliferous scale larger than the bract, A;_ laxifolia Hooker has a scale equal in size to the bract, and A^_ selaginoides Don has a bract larger than the ovuliferous scale. Miller (1976) also points out that in the Pinaceae, superficial appear­ ances of fossil cones are of limited value and can be highly misleading.

Romeroites, a genus of taxodiaceous cones from the

Neuquen province of Argentina (Spegazzini, 1924) is a structurally preserved cone described, from- a single cone specimen and some fragments. The dimensions and external appearance are very similar to those of Pararaucaria, prompting Florin (1940; 1944) to suggest that the two might be identical. Internal structure reveals, however, many- seeded ovuliferous scales, orthotropous seeds, and a bract and scale that are reportedly fused for a greater part of their length.

Cone vascularization interpreted by Wieland as being like Pinus, and by Calder (1953) as being like the

4 • ■ Pinaceae, Araucaria (section Bunya), and some Taxodiaceae was never really studied in detail. In Pararaucaria, the ovuliferous scale and bract traces are separate at their origins from the axis stele. The general configuration 33

is unlike that of Pinus and certain Pltyostrobus but is

similar to Pseudoaraucaria and other genera of the

Pinaceae. The ovuliferous scale trace has the abaxially

concave configuration like that of a pinaceous cone

(Miller, 1976, Fig. 1) and unlike that of taxodiaceous or

araucarian forms (Radais, 1894; Worsdell, 1899; Aase, 1915;

Hirmer, 1936). The bract trace is stout, unlike most pinaceous genera, except Pseudoaraucaria and some Pinus

species (Miller, 1976) . It splits and becomes extensive as in many cones of- the Taxodiaceae (Eames, 1913; Radais, .

1894).

Seed numbers are similar to the Araucariaceae. The

Taxodiaceae usually have 3-5 seeds per scale, although

as few as 2 and as many as 9 seeds per scale have been reported (Bierhorst, 1971). The extant genera of the

Pinaceae are characterized by 2 seeds per scale.

Many-layered seed integuments are known to occur in

the Taxodiaceae and Cupressaceae. Pararaucaria shows some resemblance to Cupressus funebris described by Konar and Banerjee (1963) in numbers and appearance of sclero-

testal layers. The fossil CunninghamioStrobus hueberi

(Taxodiaceae) was also described as having a multi-layered

seed integument (Miller, 1975) . The holotype (USNM

#31639) was re-examined in this study and revealed a

similar structure to the integuments of Pararaucaria in • 34

longitudinal section. The sclerotesta contains a layer of

radially elongate sclereids surrounding a region of

isodiametric sclereids. The presence of two radially

elongate rows of sclereids, a feature which indicates a

later developmental stage in Cupressus and other genera, was observed in one seed. Cunninghamiostrobus seeds do

not, however, show a ribbed integument or prominent

scl-ereid groupings.

The general shape of the seed itself is similar to

pinaceous types. The distal extent^of the wing is not

known for certain since many cones of Pararaucaria have

undergone some weathering; however, the maximum seed wing

length appears to have been approximately 2 mm. The one

character which Calder felt might indicate affinities for

Pararaucaria outside the Taxodiaceae is seed development.

Taxodiaceous seeds are free throughout their development

from the fertile scale except in the basal region (Pilger,

1926). Pinaceous seeds are not free during development,

their wings being derived from ovuliferous scale tissue.

Most Pararaucaria seeds studied by Calder (1.953) and

during this investigation are separated from the scale ex­

cept at the chalaza.. The similarity of the seed wings to

the upper scale surface, however, and the attachment of the

wings in some cones to the ovuliferous scale would indicate

that developmentally, the wing probably arose from tjie ovuli­

ferous scale tissue. Calder (1953) believed if Pararaucaria 35 could be shown to have pinaceous seed-wing development that it should be placed within a new family intermediate between the Pinaceae and the Taxodiaceae.

Structure of the seed wings is similar to that of pinaceous genera such as Abies. The anatomosing stellate sclereids which Calder (1953) compares to the "cellules etoilees" of Abies nordmanniana described by Radais

(1894) are here interpreted as glandular hair-like tri- chomes similar to those found in Cedrus by Chowdhury

(1961), which also occur on the surfaces of the ovuli­ ferous scale and from which the wings are derived. They may have functioned in cone closure after pollination by the production of resinous substances as they do in

Cedrus. Of the 20 taxa of pinaceous seed wings figured by Von Tubeuf (1892) only Pinus wallachiana A.B. Jackson

(=P. excelsa Wallich) and Cedrus libani A. Richard showed cells of the seed wing that appear to anastomose and to contain some intercellular spaces. A similar structure occurs in Abies grandis seed wings (Fig. 12). Seed wings of other species have cells with straight or1 wavy mar­ gins and no intercellular spaces.

The structure of the embryo is similar to that of other living conifers at the telo-stage period of develop­ ment (Schopf, 1943). No leaf primordia have been observed in shoot apices as occurs in Cedrus (Buchholz and Old, 1933). All the typical features of a conifer embryo at % this developmental stage are present: shoot apex, cotyle­

dons, root meristem, hypocotyl, and calyptroperiblem or

"root cap". Cotyledon number is similar to that in the

Pinaceae and Taxodiaceae. The most common number of coty­

ledons is 8 in Pararaucaria? on the other hand, 6 cotyledons

are. present in some embryos. The transverse section in

Fig. 18 is the cotyledonary node of an embryo with six cotyledons. The three segments or lobes of embryo tissue represent the number of cotyledonary primordia in Pararau­

caria. . Lobing at the cotyledonary node is known to indi­

cate the numbers of cotyledonary primordia, for example

in Larix (Schopf, 1943? Fig. 73, 77). Three or four lobes are common in Pararaucaria. In Buchholz's (1919) studies of pinaceous embryos, he found that several genera show a fusion of cotyledonary primordia to give the numbers of

cotyledons seen in the mature state. Many showed no

cotyledonary fusion? and none showed the splitting of

cotyledonary primordia to give the. number of’cotyledons

seen in the mature embryo. In Pararaucaria, on the other

hand, the three cotyledonary lobes probably represent

the number of cotyledonary primordia, that upon dichoto­ mizing resulted in the six cotyledons seen in such embryos.

Thus Pararaucaria is different than Buchholz's pinaceous

genera in this feature. Cotyledon assymmetry observed 37 at the cotyledonary node in Pararaucaria (Fig. 18) is the common case in embryos of the Pinaceae (Buchholz,

1919).

Polycotyledony in Pararaucaria is noteworthy, since this condition for conifer embryos existed concomitant with the dicotyledonous condition in Araucaria mirabilis

(Stockey, 1975a). The suggestion that polycotyledony in conifers gave rise to the dicotyledonous condition

{Buchholz, 1919) will have to' be re-examined in fossils from sediments older than the Jurassic Cerro Cuadrado deposits.

Of some interest are the two triangular shaped strands of sclerenchyma arising in the cortex of the cone axis of Pararaucaria that accompany the ovuliferous scale trace on its adaxial face into the distal part of the scale. Miller (1970) reports a dense tissue of sclereids, filling the gap between the traces of the bract and scale, arid following the traces out into the cone scale complex of Picea diettertiana. Extending through this sclerenchyma are 2 resin canals. The number varies from

2-13 in the genus Picea. There are no resin canals associated with the sclerenchyma of Pararaucaria. Their peculiar triangular configuration and persistance in the scale is unknown in other conifer genera. Calder (1953) suggests that these sclerenchyma strands may have been 38 associated with the separation of the scales at maturity and the liberation of seeds. Thus in Pararaucaria we not only see evidence of a mechanism for cone closure after * ♦ pollination by the glandular hairs of the scale surface but also a possible mechanism of cone opening after matura- l tion to effect seed release. Pararaucaria probably re­ tained its scales and shed its seeds at maturity like members of the Pinaceae and Taxodiaceae. Evidence that the seeds’ were in fact released from the scale can be seen directly in the presence of thick, woody bract and scale traces without abscission layers and an abscission layer at the seed chalaza, and indirectly by the absence of

isolated cone scales at the Cerro Cuadrado localities.m » ' Fossil ovulate cones of Pararaucaria patogonica do not conform to any extinct of extant conifer genus but instead combine a number of features from 4 or 5 families of conifers. The one-seeded condition of araucarians is almost the only feature linking it with that group; and

Pararaucaria represents a misnomer suggesting a relation­ ship that probably does not exist (Gothan, 1950).

The systematic position of Pararaucaria would also appear to be more Pinaceae centered than has been thought by previous workers. This idea corresponds partially to what Wieland (1935) originally suggested. Pinaceous characters include vascularization of the cone-scale 39 ' complex, seed wing features, cotyledon numbers, and secon­ dary xylem pitting. Seed integuments are taxodiaceous and cupressaceous. Taxodiaceous characters also include the conspicuous, woody bract and strongly developed bract trace, cotyledon numbers pith structure, and seco'n- diary xylem lacking resin canals. Pararaucaria is a petrifaction genus, therefore comparisons with the Mesozoic

Cheirolepidaceae (Hirmerella and Tomaxellia) are limited. .

This study supports Calder's idea that seed number in fossil cones appears to have a limited diagnostic value and that of Miller (1976) that external resemblances of fossil cones are often misleading. Takhtajan (1953) sug­ gests that the Taxodiaceae may have arisen from the earliest primitive Pinaceae. The intermediate position of Pararau­ caria between the Taxodiaceae and the Pinaceae may lend support to this interpretation. It may, however, suggest the opposite, i.e., the Pinaceae may have arisen from primitive taxodiaceous stock. It is the opinion of this writer that the genus Pararaucaria is sufficiently dif­ ferent in the combination of characters to warrant family status, and that further investigations of Jurassic age conifers will provide the opportunity of placing Pararau­ caria in the proper perspective with reference to other conifer families. The excellent silica permineralization of Pararau­ caria which has allowed developmental interpretations of

A. mirabilis cone structures from the Cerro Cuadrado localities (Stockey, 1975b) and the large numbers of specimens available have provided us with information about the reproductive strategies of the plant such as the retention of scale, shedding of seeds, possible mechanism for seed dispersal, and mechanism for cone closure after pollination. The polycotyledonous condition is seen to have existed in Pararaucaria during the Juras­ sic, and probably came about by the dissection of cotyledonary primordia. Preservation of the delicate embryo tissues, so rare in the fossil record, has allowed direct comparison of the telo-stage embryo of

Pararaucaria to those of living conifers at a comparable developmental stage. SEEDLINGS

Introduction

Among the remains of coniferous cones and twigs at the Cerro Cuadrado localities are a number of tuber­ like, root-like and corm-like structures of various shapes and sizes. These structures were first described by

Wehrfeld (1935) as "araucarian fossils" with no remarks as to their morphological nature or structure. In 1950,

Gothan described "fig-like bulbs" from the Cerro Cuadrado collections in Berlin and Stockholm. He examined the * specimens using reflected light and made a few thiri sec­ tions of a corm-like form for anatomical study. .Although these specimens were described as decorticated coniferous shoots lacking resin canals, the author suggests they might represent insect or fungal stem galls (Gothan,

1950).

Wieland in his monograph on the petrified forest figures 2 specimens that he believed'to be young seedlings

(Wieland, 1935j PI. 13, Fig. 2). These structures, that are reinvestigated in this study, show little external

41 resemblance to the £ig-like structures of Gothan and

Wehrfeld. Calder (1953) also suggested that the corm- like structures in the British Museum collection were possible seedlings, ‘ but her sections showed no reliable internal diagnostic features.

It is the purpose of this investigation to deter­ mine the morphological nature of a wide variety of bulb­ like or fig-like structures that have been thought to represent seedlings. These structures are compared to seedlings of the Araucariaceae and other conifer families, and their significance with respect to the reproductive biology of the plant is discussed.

FOSSIL SEEDLINGS

There are several small woody structures associated with the Cerro Cuadrado remains that resemble araucarian seedlings of the sections- Bunya or Columbea of the genus

Araucaria. These specimens range from top-shaped to turbinate (Fig. 21, 24, 25). They range in size from 3.5 -

5.3 cm. long and up to 2.5 cm. in diameter. The surface is typically smooth but may display longitudinally directed ridges or striations.

Figure 22 is a transverse section at the level indicated by A in Fig. 24. Numerous cracks that parallel 43 the long axis of the specimens extend several mm into the cortical regions of the axis. These interruptions in the tissue constitute artifacts of weathering or preservation rather than features of the tissue system. In most of the specimens there is no epidermis or periderm present.

The majority of the specimen consists of isodiametric cells that constitute a parenchymatous cortex (Fig. 22, 23, 27).

Scattered throughout the cortex are dark patches that may constitute plugged resin canals (Fig. 23). The presence of resin canals has been reported in living embryos and seedings of A^ angustifolia (Burlingame, 1915; Hill and de Fraine, 1909a) and from fossil embryos of A^_ mirabilis from this locality (Stockey, 1975a). It is interesting to point out'that resin canals in living araucarian seedlings show a more regular distribution than those observed to date in the fossil material.

The central portion of the axis contains a pith com­ posed of cells that are of slightly larger diameter than those of the cortex (Fig. 26). Radial rows of small cells extend out from the central pith- region and consti­ tute the vascular tissue of the axis (Fig. 22, 23, 25).

Preservation of these cells has made it impossible to de­ termine the maturation pattern of the primary body or the pattern of pitting on the tracheid walls. At lower levels 44

(Pig. 24-B) of the seedling, the vascular system is com­

posed of an elliptically shaped xylem cylinder.

LIVING ARAUCARIAN SEEDLINGS

Detailed studies delimiting the anatomy of extant

araucarian seedlings are few. In 1906, Seward and Ford

discussed differences in seedling morphology and structure

in two of the sections (Eutacta and Columbea) of the genus

Araucaria. Detailed anatomical information including the * * organization of the vascular system.at different levels

of the seedling for A^_ anqustifolia (=A. brasiliensis),

A. cunninghamii (Hill and de Fraine, 1909a), and A^ bid-

willii (Shaw, 1909) provided the most comprehensive treat­

ment of seedling anatomy. Information on morphological

features relative to the of araucarian

seedlings is found in contributions of Durer (1865),

Blanchard (1892), Heckel (1892) and Hemsley (1901).

Seedlings of the section Eutacta are characterized

by , slender axes, and 4 linear

cotyledons (Seward and Ford, 1906; Wilde and Eames, 1952).

Figure 28 shows a germination sequence of A^ heterophylla

(Salisbury) Franco (=A. excelsa (Lambert) R. Brown), in­

cluding radicle emergence, cotyledon emergence, hypocotyl

extension, cotyledon separation, and branch initiation. 45

There are 4 cotyledons present which are photosynthetic

t and remain on the plant for at least 1 year. Hill and de

Fraine (1909a) state that seedlings of the Eutacta section have 2 cotyledons which fork above the cotyledonary node giving the illusion of 4 cotyledons. The external morpho­ logy of these seedlings is the typical conifer type

(e.g., Pinus). There is a striking resemblance between all species of this section in the appearance of the ju­ venile foliage even though some of the adult foliage shows distinct morphological differences. In none of the seed­ lings in this section are there any hypocotyl swellings.

Seedlings of the sections Columbea and Bunya, on the other hand, have been reported as having , pronounced hypocotyl swelling and two cotyledons which are often fused at their bases.(Seward and Ford, 1906; Shaw, 1909; Hill and de Fraine, 1909a).

Araucaria angustifolia shows a typical seedling structure of the section Columbea (Fig. 29, 32) grown under green­ house conditions. The 2 cotyledons of these seedlings remain embedded in megagametophyte tissue for many months after germination (Fig. 29). They are often fused into a tube at the cotyledonary node and surround the epicotyl.

Beneath the cotyledonary node, the hypocotyl of the seed­ ling shows varying degrees of swelling (Fig. 29, 32).

Hypocotyl swelling may be a more pronounced feature in nature where the seedling must undergo dormant periods due to moisture changes than those grown under more favorable greenhouse conditions.

A transverse section of a.seedling of A^ angustifolia in the upper portion of the swollen hypocotyl shows a massive region of cortical parenchyma cells packed with numerous starch grains (Pig. 37). The outer portion of the axis consists of a narrow band of periderm 5-6 cells wide that.arises just outside the outer ring of resin canals

(Fig. 37). In transverse section, resin canals occur in two concentric rings in the ground tissue of the axis, one near the periphery of the axis, the other surrounding the vascular bundles (Fig. 37). Burlingame (1915) has reported a.ring of resin canals 3-4 cells beneath the epi­ dermis in dormant embryos of A^_ angustifolia and another just outside the procambial strands. A similar situation has been reported in the dormant embryos of the fossil

A. mirabilis from the Cerro Cuadrado region (Stockey,

1975a). Burlingame (1915) and Hill and de Fraine (1909a) note a small number of resin canals scattered throughout the cortex in addition to the 2 concentric rings. Figure

37 illustrates a similar disposition of resin canals in

‘this material. 47

The vascular system at this level is composed of

4 concentric bundles {Fig. 37) . Each bundle consists of a small lens of endarch primary xylem from which radiate up to 11 rows of tracheids of the secondary xylem (Fig. 30). Thin-walled cells of a vascular cam­ bium separate sieve cells and sclerencyma fibers of the

t secondary phloem and crushed primary phloem elements from the xylary tissue (Fig. 30). . Transition in xylem con­ figuration takes place in the lower portion of the swollen hypocotyl with a reduction to the diarch condition.

Hill and de Fraine (1909a) also report diarch lateral roots.

The pith region is composed of a similar appearing ground tissue to that found in the cortical zone. No resin canals have been observed in the pith of Araucaria angus­ tifolia.

The seedling structure of A^ bidwillii is similar to that of A;_ angustifolia with respect to mode of germination and the presence of hypocotyl swelling. The degree of swelling of the Bunya seedling, however, is more pro­ nounced (Fig. 31, 36, 38) . After the radicle of the A. bidwillii seedling emerges, it grows downward concurrent with the swelling of the hypocotyl (Fig. 38). The cotyle­ dons surround the shoot apex and do not rise above the 48

ground on germination (Fig. 36, 38). After approximately

4 months, the cotyledons eventually break away and the

shoot emerges from the ground. Hemsley (1901) reports

that the seedlings show 2 phases of growth beginning with

the formation of a "carrot-shaped woody body". Once this

stage of development is reached, seedlings, may withstand

a dormant period of considerable length without vitality

. loss. Seward and Ford (1906) report Tidmarsh's success

in germinating the "wooden carrots" recovered from a

wrecked vessel, but also expressed the opinion that this

interval was not necessary to continued successful growth.

At the present time, gardners at the forestry station in

Imbil, Queensland germinate the seedlings to the stage seen

in Fig. 38. Seedlings at this stage of development are

then harvested and transplanted with the shoot exposed

above the ground. This stimulates the initiation of the

second phase of growth involving foliar development.

Representative transverse sections of an Araucaria

bidwillii seedling (Fig. 36) are illustrated in Figures

33 and 34. Embedded within a parenchymatous ground tissue

containing numerous starch deposits (level A; Fig. 36)

are vascular elements arranged in two concentric zones

(Fig. 33). The cotyledonary bundles of the outer zone

(Fig. 33) are variable in number and general morphology

whereas the inner fused ring of plumular bundles is 49 composed principally of secondary vascular elements. In the region of the cotyledonary node, resin canals are arranged in a ring just beneath the periderm. Numerous branching resin canals are also present within the cor- test between the cotyledonary*bundles.

At the approximate level B (Fig. 36), the vascular system consists of a single ring of discrete bundles

(Fig. 34). Each vascular strand is histologically similar to the bundles described for A^ angustifolia (Fig. 30).

All seedlings at this stage of development possess a narrow band of periderm 5-6 cells thick that arises just beneath the epidermis (Fig. 34). Resin canal dis- tribution and morphology parallels that described for

A. angustifolia (Fig. 37). At the most distal level of the seedling, the primary root assumes a diarch configura­ tion.

OTHER FOSSIL SEEDLINGS

Based upon external features alone, Wieland (.1935) regarded the two fossil specimens illustrated in Figures

35 and 39 as 2 and 3 year old seedlings. Anatomically, however, it is clear that the axis represents a decorti­ cated stem with abundant secondary xylem composed of slender tracheids and uniseriate rays, and lacking resin canals (Fig. 41, 42). The pith is composed o£ large (65um) cells, some of which contain dark contents. Square to rectangular sclereids characteristic of cone axes of

Pararaucaria patagonica are scattered throughout the pith.

Small patches of periderm are present on the external sur­ faces of the two axes (Fig. 35).

In tangential section, the wood shows uniseriate rays 2-6 cells high (Fig. 41).. Pitting consists of uniseriate circular-bordered pits. Anatomically this wood resembles cone axes of Araucaria mirabilis and P^_ pata­ gonica from this locality.

There are some indications of the presence of growth rings in Wieland's two seedlings (Fig. 42) . These are generally discontinuous but probably indicate 2-3 year development since there are at least 4 growth increments in some portions of.the axis (Fig. 42).

Sections made of the root-like appendages that arise from the base of the seedlings demonstrate a similar anatomy to stems near the axis and possess a diarch stele in the main root. The diarch organization is similar to that in A^ bidwillii (Shaw, 1909), angustifolia (Hill and de Fraine, 1909a), A^ araucana (Seward and Ford, 1906), and other conifer seedlings with two cotyledons (Hill and de Fraine, 1908). 51

In addition to the seedling structures already des­ cribed, there are several corm-like structures in the Cerro

Cuadrado collections (Fig. 40, 43-52). These curious structures are similar to those first figured by Wehrfeld

% (1935) and subsequently studied by Gothan (1950) and

Calder (1953). A transverse section midway through the seedling in Figures 40, 43 and,44 shows the typical struc­ ture of the corm-like 'forms (Fig. 53). There is a central pith which is spherical to oval in outline containing isodiametric cells surrounded by a massive area of secon­ dary xylem lacking resin canals. In longitudinal section, there is little structural detail except in the better preserved elongate cells near the central portion and in areas where the vascular rays extend through the tissues.

One partially decorticated specimen (Fig. 56) shows a continuation of the main axis into a root-like structure and the presence of a lateral branch or secondary root.

A tangential section of the corm-like structure shows the lateral axis in transverse section (Fig. 55). The rem­ nants of periderm and other extra-xylary tissues including the cambium and phloem appear to be present. The corm-like expanded portion of the structure shows some vascular rays

(Fig. 55, arrows).

The two elongate structures illustrated in Figures

54 and 57 superficially resemble portions of seedlings such 52 as those described earlier (Fig. 21, 24, 25). Anatomically,, however, these structures are most similar to Wieland's two seedlings as well as the ubiquitous wood of Araucarites santaecrucis Calder (1953) from the Cerro Cuadrado, and consist mostly of secondary wood. Even though there is a morphological external tapering to these specimens, the width of the pith region remains constant for the length of the specimens, indicating that the gradual taper is the probable result of differential decortication of shoot systems.

DISCUSSION

There are a large number of specimens referred to as seedlings present in the Cerro Cuadrado collections displaying a variety of morphological and anatomical characters that probably represent a number of different plant structures. The seedlings illustrated in Figures

21, 24 and 25 show a similarity to first year seedlings of Araucaria sections Bunya and Columbea, and unlike those, of the section Eutacta. The massive cortical zone with isodiametric parenchyma cells and the reduction of the stele to a diarch condition in the root appear to indi-' cate affinities with the genus Araucaria. The presence of a diarch stele suggests the seedlings may have had 53

2 cotyledons. Hill and de Fraine (1908, 1909a) have shown that seedlings in the Cupressaceae, Cephalotaxaceae,

Taxaceae, Taxodiaceae, and Araucariaceae that have 2 cotyledons possess diarch roots. Other gynrnosperms do not follow this type of pattern of organization (e.g..

Ginkgo, Cycadales, and most of the Pinaceae (Hill and de

Fraine, 1909a, b).

The presence of the swollen hypocotyl in gymnosperm seedlings is rare, apparently confined only to the sec­ tions Columbea and Bunya of the genus Araucaria, and

Encephalartos and Stangeria in the Cycadales. Other mem­ bers of the Cycadales, e.g., Zamia, show enlarged roots, but not the pronounced hypocotyl swelling.

It is my opinion that the structures illustrated in Figures 21, 24 and 25 are seedlings of the Araucariaceae, sections Bunya or Columbea, exhibiting the typical swollen hypocotyl of this group, and by inference probably pos­ sessed hypogeal germination. The relationship of these seedlings compares most closely with the seed cone of A. mirabilis, since this taxon has two cotyledons per embryo, and the type of embryo associated with hypogeal germina­ tion. In addition, A^ mirabilis shows many similarities to the section Bunya of the genus Araucaria (Darrow, 1936;

Calder, 1953; Stockey, 1975a), a section containing A. bidwillii, the species that shows the most prominent 54 hypocotyl swelling. It is also suggested that they repre­ sent first-year seedlings due to their lack of abundant secondary tissues and the similarity in structure to first- year seedlings of Araucaria bidwillii and A^ angustifolia

{Columbea). Pararaucaria patagonica, the other taxon of conifers found at the Cerro Cuadrado.Petrified Forest, has 6-8 cotyledons per embryo. In addition, this taxon has no relationship to the Araucariaceae, but probably has affinities with the Pinaceae and Taxodiaceae, families in which there is no hypogeal germination or hypocotyl swel­ ling.

The seedlings illustrated by Wieland (1935) do not show pronounced hypocotyl swellings. The structure of the wood is similar to cone axes of both A. mirabilis and P. patagonica. In the pith, however, the presence of large cells with dark contents and rectangular sclereids would favor affinities with P_;_ patagonica, and they probably represent the seedling structures belonging to this taxon.

Since Pararaucaria is believed to have pinaceous affini­ ties, the presence of a diarch root and absence of hypo­ cotyl swellings agree with the observations of Hill and de Fraine (1909a, b) with respect to seedling structure in most members of the Pinaceae.

Cones of Aj_ mirabilis are more numerous in the Cerro t Cuadrado collections than those of P^ patagonica. Whether 55 this is due to the larger size of the specimens and some undetermined bias interjected during collecting, or merely reflects the abundance of the taxon in the forest, is presently not known. A^ mirabilis, however, was probably the most likely species from the locality to demonstrate forest regeneration. Cones of patagonica, even when apparently mature, rarely show the presence of embryos within the seeds (only about 5%). When the cones have been fertilized, about half of the seeds possess embryos.

A. mirabilis, oh the other hand, shows a larger number of cones with seeds containing embryos (25-35%). When cones contain seeds with embryos, about 75-85% of the seeds are fertile. Such evidence, while indirect, does substantiate the widespread occurrence of Araucaria mirabilis as a dominant element of the Cerro Cuadrado forest during the

Jurassic Period.

It is my opinion that the woody structures in

Figures 54 and 57 represent decorticated shoots. This interpretation was originally proposed for all of the

Cerro Cuadrado seedlings by Gothan (1950). The constancy of the diameter of the pith throughout the length of the axes and the similarity of these structures to the wood of

Araucarites santaecrucis lend support to this interpreta­ tion . 56

The woody corm-like structures (Pig. 40, 43-52) are more'difficult to accurately position with reference to origin and parent group of organisms. They do, how­ ever, show a gradation from turbinate (Fig. 40, 43, 44,

56) to a more flattened fig-like or bulb-like form

(Pig. 45-52). These structures are woody, appear to be constructed entirely of secondary xylem lacking resin canals, and are usually decorticated (Fig. 53) . There is a central pith-like region that changes from spherical to oval in outline. Wehrfeld's. (1935) material was almost entirely of this type. Gothan (1950) figures similar appearing structures. The various sections prepared by

Gothan are anatomically identical to the specimens illustrated in Figures 53 and 56. He did not rule out the possibility, however, that these were fungal stem galls or'insect galls. The presence of secondary xylem negated his initial assumption that they were plant exu­ dates.

In Darwin's Journal of Researches (1839, p. 298-

99) , he describes a fungus parasitic on . in

Tierra del Fuego. There is a superficial resemblance of this ascomycete to the turbinate and corm-like forms.

i There is, however, no evidence of a hymenium or any identi­ fiable fungal hyphae in the Cerro Cuadrado material as would also be apparent in this type of ascomycete or in 57 the fungal stem galls. Nor are there any irregular growths of tissue as in insect galls. An insect entering a stem usually feeds upon the pith, vascular rays and primary xylem. Its presence will cause cambial activity to in­ crease markedly, but there is also a proliferation of parenchymatous tissue in the larval cavity, eventually leading to a loss of symmetry in the vascularization

{Allen, 1951; Beck, 1954). There is no loss of symmetry or tissue proliferation in any of the seedling structures examined.

The regular arrangement of tracheids throughout these structures suggests affinities with older decorti­ cated seedlings. It is interesting to note that in later seedling development, A;_ bidwillii depletes much of the stored starch in the hypocotyl concomitant with an in­ crease in secondary xylem formation.

The small number of conifer taxa and large number of seedling specimens found at the Cerro Cuadrado locality support the conclusion that the turbinate and corm-like structures do in fact represent different stages of seedling development. Their similarity in structure to living seedlings in the sections Bunya and Columbea of the genus Araucaria is reflected in both anatomical and morpho­ logical detail. The occurrence of the rare swollen hypocotyl and hypogeal germination in conifer seedling 58 development appears to have been present by at least

Jurassic time. The establishment of systematic parameters for all types of fossil developmental stages provides an opportunity to more accurately investigate not only the total biology of the organism, but of equal importance, • provides a basis for understanding the evolution of more subtle biological phenomena such as germination sequence and various embryological stages'. Lastly, the charac­ terization of all developmental phases of several Jurassic age conifers provides the only available method whereby early conifer phylogeny may be fully and accurately elucidated. ARAUCARIA MIRABILIS (SPEGAZZINI) WINDHAUSEN

EMEND. CALDER

Introduction

. The silicified ovuliferous cone Araucaria mirabilis, the larger of the two cones found at Cerro Cuadrado, was first described by Spegazzini (1924) as Araucarites mira­ bilis. A year later Gothan (1925) published the name

Araucarites windhauseni for cones which he had received from Windhausen's collection. Wieland (1929) published

2 tentative names, Proaraucaria mirabilis and Proaraucaria elongata for these cones and a third name, Proaraucaria patagonica, for what he thought was a mature staminate or microsporangiate cone. None of these names were validly published since none of the cones were illustrated.

However, after further examination of the cones, Wieland

(1935) placed Spegazzini's Araucarites mirabilis, Gothan's

Araucarites windhauseni and his own Proaraucaria elongata into synonomy as Proaraucaria mirabilis (Wieland) which he had retained as the preferred combination. Wieland thought the range of variation exhibited was not sufficient

59 60

to distinguish two species; he did, however, make a dis­

tinction on size by creating^a new variety. Proaraucaria mirabilis var. elongata based on his .1929 description.

Proaraucaria minima sp. nov. was illustrated in Wieland's plates, but had no accompanying description and is there­

fore regarded as nomen invalidum. He did, however, pro­

vide a description for a taxon Proaraucaria mirabilis var. minima in the text. He also states that there is only a

single species of Proaraucaria and a couple of varieties;

therefore, Proaraucaria minima sp. nov. in the plate ex­

planation was probably an unintentional mistake on his

part and is not considered a new species.

Calder (1953) amended the diagnosis of Proaraucaria

and placed all of Wieland's species and varieties under

the binomial Araucaria mirabilis, concluding there were

no anatomical differences to separate these taxa from each

other. Calder further notes that the generic name Arau­

caria should be used with Araucarites reserved for cones

or branches with the araucarian habit which lack preserva­

tion of structural .details. Since Araucaria mirabilis is

structurally preserved, the generic name Araucaria is

used even though Araucarites has taxonomic priority.

Calder has assigned Araucaria mirabilis to the section

Bunya Wilde and Eames (1952) that contains ‘the extant

Araucaria bidwillii Hook., native to Queensland, . 61

Stockey (1975a) described in detail the tissue systems of the mature seeds and embryos of Araucaria mirabilis and

* pointed out similarities between this species and extant araucarians.

Most Araucaria mirabilis cones described from the

Cerro Cuadrado Petrified Forest are large, ranging about

6-10 cm in length and diameter. However, there are a num­ ber of small cone specimens in the Cerro Cuadrado collec­ tions that measure 3-4.5 cm in diameter, which have been suggested to be younger developmental stages of Araucaria mirabilis (Wieland, 1935; Calder, 1953; Stockey, 1975b).

Wieland (1929) was the first to suggest that a younger growth stage of A^_ mirabilis was present in the Field Mu­ seum collection, but expressed the opinion that the finer details of histology of cone tissues were not present.

Wieland (1935) named the variety Proaraucaria mirabilis var. minima for some of these supposed younger cones. He also provided a description for the new species Proaraucaria patagonica which .he felt was a mature "staminate" or microsporangiate cone. He, however, expressed the opinion that this specimen (P13981, Plate 6, Fig. E) might actually prove to be a younger seed cone. Calder (1953) believed all of Wieland's (1935) species probably represented dif­ ferent developmental stages of the same cone taxon. She further suggested that one specimen (P13981) was actually a naked cone* axis, since it had a similar appearance to . cone axes of extant Araucaria species..

It is the purpose of this' investigation to provide a more .detailed analysis of many of the supposed younger cone developmental stages, the Araucaria mirabilis cone axis, and the "staminate" cones of Wieland. The reproduc­ tive mechanisms and development of Araucaria mirabilis are elucidated. Development- and reproductive biology are compared with extant genera in the Araucariaceae and other conifer families.

Araucaria mirabilis (Spegazzini) Windhausen emend. Calder

• i SYNONOMY

1924. Araucarites mirabilis Spegazzini p. 126 text-figs. 1-3, 4 (1-7).

1925. Araucaria windhauseni Gothan p. 200, pll 2-7, pi. 8, fig. 1.

1929. Proaraucaria mirabilis (Speg.) Wieland p. 60, nomen invalidum (no published figure).

1929. Proaraucaria elongata Wieland p. 60, nomen invali­ dum (no published figure).

1931. Araucaria mirabilis (Speg.) Windhausen p. 201.

1935. Proaraucaria mirabilis (Speg.) Wieland p. 19, pi. 1, pi. 7, figs. 2.,3; pi. 10, pi. 11, figs. 1,2,4, text-fig. 4.■

1935. Proaraucaria mirabilis var. elongata Wieland, p. 26, pi. 8, fig. 1? pi. 9, fig.‘2; pi. 12, fig. 4, 63

1935.' Proaraucaria patagonica Wieland p. 26, pi. 6, £igS. B,D,E,F,.

1935. Proaraucaria minima Wieland, p. 174, pi.. 6, fig. A, • p^ 176, pi. ITT, fig. 3.

1935. Proaraucaria mirabilis var. minima Wieland, p. 26, pi. 6, fig. A; pi. 12, fig. 3.

1953. Araucaria mirabilis (Speg.) Windhausen emend. Calder p. 110, pi. 3-5, text-figs. 2-4.

Emended Diagnosis

Araucaria mirabilis (Speg.) Windhausen — Ovulate cones, spherical to ellipsoid in shape, 2.5-10 cm in length,

2.5-10 cm in diam. Many cone-scales, spirally arranged, one seeded. Bract woody, with rhomboidal apophysis and deciduous laminar tip. Ovuliferous scale fused to bract upper surface for 1/2 length. Scale approximately 0.8-2.3

cm long and 0.5-1.5 cm wide including wing-like bract

extensions. Ovuliferous scale about 0.5-0.9 cm long,

0.4-0.5 cm wide, and 0.1-0.2 cm high. Seed approximately

0.8-1.3 cm long, 0.2-0.6 cm wide, anatropous, wingless, embedded in ovuliferous scale. Three-layered integument,

sclerotesta ranging from 355-715 urn thick, tightly packed

sclereids in zig zag pattern on seed surface at maturity.

Nucellus free except at base, wavy outline at micropylar

end, extending out of micropyle in younger seeds. Mega-

gametophyte extensive, cells isodiametric; archegonia. 64

sunken, borne singly near micropylar end. Embryo dicotyle­ donous, about 2 mm long, 0.25 mm wide and deep, 6-8 vas­ cular strands to each cotyledon; ring of 10-12 resin canals beneath protoderm of cotyledons and hypocotyl.

Cone axis approximately 0.8-1.2 cm in diam.; pedun­ cles up to 1.1 cm long. Wide pith with branched sclereids surrounded by ring of fused vascular bundles at base that split at higher levels in the cone. Cortex with resin canals and fibers; phloem and cambium usually crushed.

Two traces per cone-scale; bract trace single, scale trace i formed by fusion of 2 inverted traces; both branch distally, accompanied by a series of resin canals; ligule strongly vascularized. Seed supplied with numerous vascular bundles.

DESCRIPTION

Cone structure — The cones illustrated in Figs. 58-

66 are representative of the smaller group of Araucaria mirabilis cones in the Cerro Cuadrado collections. They range in size from 2.5-4.5 cm long to 2.5-4.0 cm in diam.

and are spherical to ellipsoid in shape. Cone-scale com­ plexes are spirally arranged and small, 4-7 mm wide' and

3-4 mm thick. Some specimens show the remains of the deciduous laminar tip (Fig. 58). Calder (1953) illus­

trates what whe terms an absciss zone on the laminar tip 65 of the bract. She points out that the surface morphology of the cones varies according to age, condition of the conev or degree of weathering. In more weathered specimens, the ovuliferous scale tip (Fig. 60) or seeds are exposed

(Fig. 63). One specimen in the British Museum collections

(V.31414) consists of five small cones in this size range that are embedded in the volcanic ash matrix and which t display an excellent state of • preservation (Fig. 61, 65).

Cone peduncles that are attached to many of the specimens

(Fig. 58, 59, 61, 64) reach, a length of 3 mm and width of 7 mm,.and show no leaf scars externally. The cones described in Wieland*s monograph as Proaraucaria patagoni­ ca, "staminate" forms were reexamined during this study and have been subsequently placed with the suite of young ovulate cones of Araucaria mirabilis (Wieland, 1935; pi.

6, fig. B,C,D).

Figures 67-70 and 72 illustrate cones of intermediate size range (about 4.5-6.0 cm in length and diam.) that were initially thought to be intermediate in development between the two groups of cones, the small cones in

Figures 58-66 and those mature forms described earlier

(Stockey, 1975a). Cone-scale complexes range in size from

7-9 mm wide and from 3-5 mm thick. External appearances are similar in most instances to those of the mature cone of A. mirabilis. 66

Cone axis — The pith of the cone axis of Araucaria mirabilis is typically composed of parenchymatous' cells and numerous sclereids (Fig. 71), and in the mature cones measures 0.4-0.6 cm (Stockey, 1975a). The vascular system of the cone axis consists of a continuous ring of fused bundles. At higher levels in the axis, these bundles be­ come separated into a large number of discrete strands.

The vascular supply to the cone-scale complex is double as

its origin (Fig. 73) as in cones of A^ bidwillii (Wilde

and Eames, 1948). These traces branch within the cortex

forming an upper and lower series of vascular strands which pass out into the bract and ovuliferous scales. Pitting of -these elements is spiral to circular-bordered. Trans­

fusion tissue occurs between the vascular strands of the bract and ovuliferous scale.

There are a large number of protoxylem points in

A. mirabilis cone axes when compared with those of

Pararaucaria from this locality (Fig. 71). Maturation of

the primary body is endarch and primary xylem pitting is

spiral. Pitting of the secondary xylem is uniseriate to

biseriate (Fig. 74). Outside the secondary xylem, the

regions of cambium and phloem are not usually well pre­

served (Fig. 71) however, numerous sclerenchyma fibers

are seen in the phloem and cortical regions as occur in

living araucarian cone axes (Thomson, 1913) . 67

Resin canals — Resin canals are numerous through­ out most of the tissues of JW mirabilis cones with excep­ tion of the xylem (Fig. 71, 75-83). These canals are usually lined with an epithelium that contains glandular hairs in the outer portions of the cone-scale complex

(Fig. 75,77). Younger cones that have not become as ex­ panded or lignified as the larger mature types provide a better system in.which to study their course throughout

i the cone tissues (Fig. 76, 79-81). The cone axis contains

2 rows of resin canals, one in the inner, the other in the outer cortex (Fig. 71). . Thomson (1913) has also reported two series of canals in the cones axes of . Although Aj_ araucana does not have canals in the pith of stems or seedlings, it does have resin canals in the pith of the cone axes (Thomson, 1913) . No distinct resin canals have been observed in A^ mirabilis cone axes.

The resin canals of the outer cortex connect with the series of canals found in the ovuliferous scales and bracts. Fig. 76 shows a tangential section of a younger

A. mirabilis cone in which as many as 10-12 resin canals supply the bract. These canals, as in the mature cones

(Stockey, 1975a), underly the vascular traces to the woody winged bract. Seward and Ford (1906) report that the resin canal distribution in araucarian cone-scales closely parallels that seen in the leaf of a particular 68 species. For example, canals in cones of Agathis austra­ lis alternate with vascular bundles as in the leaves. In

Araucaria rulei, on the other hand, they lie below each vein. If this trend in resin canal placement is reliable, one would expect leaves associated with mirabilis twigs to have resin canals underlying.the vascular strands.

In addition to the resin canal system in the bract, there is an extensive system within the ovuliferous scale

(Fig. 79,80). Up to 12 resin canals have been observed in the most distal portions of the "ligule" (free tip of the ovuliferous scale). The integuments of young ovules of Araucaria mirabilis also exhibit a resin canal system which is.all but obliterated in mature seeds due to the expansion of the sclerotesta. Two main canals exist on either side of the seed integument in transverse sections

(Fig. 79-82), forming what Wilde and Eames (1948) call the lateral ridge of the seed. Eight to ten smaller canals occur in the upper and lower portions of the ovuliferous scale in contact with the outer integument (Fig. 79,80).

At later stages, these canals are the first to be flat­ tened and obscurred by integument expansion.

Figures 81 and 82 are tangential sections of a cone measuring 4.5 x 5.5 cm which has seeds containing some smaller embryos and that probably represents an inter­ mediate between those cones illustrated in Figs. 58-66 and 69 those mature forms described earlier (Stockey, 1975a).

The two lateral resin canals of the seed are most prominent at this stage of development (Fig. 81).

Resin canals in the outer portions of the cone-scale complexes extend to the tips of the scales and may often be seen in reflected light (Fig. 78, 83, 84). They pro­ bably also extended into the deciduous laminar tip of these cones (Calder, 1953) .

Isolated cone axes — Two specimens (Fig. 87,88) were thought to represent "staminate" or microsporangiate cones (Wieland, 1935) . . Both specimens have a series of rhomboidal scars on their external surfaces borne in a. helical arrangement. Calder (1953) suggested they were naked Araucaria mirabilis cone axes, and in transverse section (Fig. 89) do exhibit the typical structure of an Araucaria mirabilis cone axis. The center of the axis has a sclerotic pith surrounded by a large number of vascular strands (Fig. 89) . In the cortical region are the two characteristic rows of resin canals (Fig. 89).

The rhomboidal scars on the surface (Fig. 87) represent the former place of attachment of the cone-scale complexes.

Further evidence that these two specimens are cone axes of Aj^ mirabilis is that not only do they show the same anatomical features and the same diameter as the cone axes, but also the presence of certain cone axes in the 70 collection that show, the cone-scales partly removed re­ vealing rhomboidal scars (Fig. 85). Figure 86 is a mature

A. mirabilis cone from which the cone axis has been removed either before or during the process of fossilization.

The cavity left when the axis is absent corresponds in size to the specimens illustrated in Figs. 87 and 88.

Cone-scale complex — The cone-scale complex of

Araucaria is typically characterized by having a large woody bract overlain by a smaller ovuliferous scale, the free portion of which has been referred to as the "ligule"

(Seward and Ford, 1906; Eames, 1913; Wilde and Eames,

1948; Thomson, 1905). The space between the ovuliferous scale and bract, termed the "ligular sulcus", has been the object of much discussion by previous workers (Wie- land, 1935; Calder, 1953). Wieland (1929, 1935) dis­ tinguished A^ mirabiiis as a new genus, Proaraucaria, based principally on this feature and the large size of the ligule. Calder (1953) did not accept these differences to be of sufficient magnitude to warrant the erection of new genus. As Wilde and Eames (1948) have pointed out, the depth of the ligular sulcus is dependent on the age of the cone; younger cones exhibit more of a separation be­ tween bract and scale. Figure 90 shows a young cone of

A. mirabilis in which the ligular sulcus is extensive, almost 2/3 of the length. Later stages .(Stockey, 1975a) 7 1 of*cone development show a sulcus about 1/3 the depth.

The cone illustrated by Wieland (1935, PI. 7, Fig. 3) with a wide sulcus' was reexamined and demonstrated to pos­ sess many abortive or sterile ovules which probably ceased growth in the first year of development (see section on abortive ovules, Fig. 97).

The ligule is strongly vascularized in A^ mirabilis simildr to that seen in A^ bidwillii (Wilde and Eames,

194B). ‘ A series of 6 vascular bundles are seen in the most distal portions of the ovuliferous scale of mature * cones. In the young cones (Fig. 58-66)/ however, vas-

r cular tissue is not readily discernible at this develop­ mental stage.

Bract tissues in the younger cone-scale complex pa­ rallel those of the scale. The most distal portions of the bract contain a series of 10-12 vascular bundles which underlie the resin canals and which continue out to

t the laminar tip (Fig. 78,83) . Tissues of the bract and ■ scale at this developmental stage are already sclerotic, a feature which increases with time (Fig. 79,80).

Ovules and seeds — Mature seeds measure 0.8-1.3 cm in length and 0.2-0.6 cm wide. There is one wingless seed per cone scale complex embedded in the ovuliferous scale tissue with the micropyle directed toward the cone axis (Fig. 91,92). The cone-scales near the cone peduncle, 72 as in living araucarian cones (Seward and Ford, 1906), do not have ovules associated with them (Fig. 91).

Ovules in the smaller mirabilis cones are 2-3 mm long and 1.5 mm in diam. and.seed integuments are in intimate contact with ovuliferous scale tissues (Fig. 79,

80)

Integuments — The integument of A^_ mirabilis ovules is three-layered (Fig. 80,81,92). Wilde and Eames (1948) report that at the pollination stage, the integuments of the ovule of A;_ bidwillii have begun differentiation into

3 layers and most of the expansion of the ovule is in a downward direction. The integument illustrated in Figure

80 is at about the stage described by Wilde and Eames

(1948) and Burlingame (1915) as the pollination stage.

In intermediate sized cones, integuments are commonly

500 urn thick (Fig. 81). Seed integuments reach up to 1.0 mm in mature specimens (Fig. 92).

The sarcotesta, or outer layer, is fused to the ovuliferous scale tissue for a greater part of its length'

(Fig. 80). It measures 70-90 urn in thickness in young ovules and is thickest in the upper half which is not restricted from expansion by the underlying bract and scale. The resin canal system of the ovuliferous scale invades the outer portions of this tissue. Numerous spherical structures 5,0-6.5 urn in diam. are associated 73

with this fleshy layer of tissue as well as with the other

integumentary layers (Pig. 80). These are similar ap­

pearing structures to those described in mature embryos

of mirabilis (Stockey, 1975a) and embryos, pith cells,

and megagametophyte of Pararaucaria, and which have been

compared to nuclear material or various storage products.

The sclerotesta or stoney layer in young ovules

which is 90 urn thick on the undersurface of the ovule

becomes almost indistinguishable from other integumentary

layers in the upper half (Fig. 80). This layer is not

yet lignified and is expanded to only about 1/5 of its

maximum thickness. Intermediate sized cones show a

sclerotesta approximately 350 um thick (Fig. 81, 82).

The sclerotesta at maturity shows a massive layer (up

to* 715 um thick) of branched sclereids in intricate zig

zag patterns that in the mature seed make up the bulk of

the seed coat in older cones (Fig. 91,. 92).

The endotesta which is also fleshy at an early stage

of development is 30-40 um in diameter in young cones and

forms the inner boundary of the seed coat (Fig. 80).

In Figures 81 and 82 the ovules show intermediate size

range between the endotesta layer in the young and the

mature A^_ mirabilis cones.

Ovule Vascularization — Vascularization of the

ovule is complex and similar in many respects to that

4 74

illustrated for JW bidwillii (Wilde and Eames, 1948).

One.to four vascular bundles have been observed within the

integument of some seeds (Fig. 93). The vascular supply

to the ovules at younger developmental stages (e.g.,

Fig. 58-66) is not sufficiently differentiated to enable a reconstruction of its course through the ovuliferous

scale as was possible with the resin canal system. Wilde and Eames (1948) report a similar situation in A^ bidwillii ovules at the pollination stage in which the ovular supply

is not sufficiently differentiated to distinguish it from other cone tissues. Cones of intermediate size (P13888,

P13935, 031221E, P13828II, SR2018), however, do show dif­

ferentiated tracheids that are not obscured by excessive cone-scale and seed integument expansion. Figure 96

illustrates the vascularization near the chalazal end of

the seed of A^ mirabilis. There is a wide area of

tracheary elements, smaller but. probably similar to the vascular plexus (Wilde and Eames, 1948) in A^ bidwillii.

Tracheids are short and have alternate multiseriate cir­

cular-bordered pits on their radial and tangential walls.

Nucellus — The araucarian nucellus which Eames

(1913) has described as "stipitate" is free from the en­

closing integuments except at the chalazal region where

it is fused to the inner layer of integument (Fig. 90,92) .

Some A. mirabilis ovules show cellularized nucellar tissue (Pig. 94) and mature seeds exhibit the papery thin wavy nucellus characteristic of mirabilis (Pig. 92). A

few cones in the 4.5-6.0 cm range show -nudellar tissue extending out of the micropylar ends of seeds (Stockey,

1975a, Pig. 12). in the cones illustrated in Figures 64 and 66, cellularized nucellus was also found extending out of the mouth-shaped micropylar opening (Fig. 94). No distinguishable pollen tubes have been seen in the nucellus of any of these cones as occurs in living species at this stage of development (Thomson, 1905a, 1907; Eames,

1913; Burlingame, 1915; Wilde and Eames, 1948; Favre-

Duchartre, 1963). The outer boundary of the elliptical shaped or tri-lobed structures in transverse sections of ovules (Fig. 79,80,95) is cellularized nucellar tissue.

This apparently plasmolyzed tissue is attached at the seed chalaza (Fig. 90). Ovules at this stage of development have been illustrated by Hirmer (1936) in Sequoiadendron giganteum (Lindley) Buchholz (=Wellingtonia gigantea

Lindley) in which the nucellus becomes wavy in outline near the apex and is separate from the integuments

for most of its length (Arnoldi, 1900; Buchholz, 1939)

and in Chamaecyparis nootkatensis (Owens and Molder, 1975,

Fig. 10) during the free nuclear stage of megagametophyte development. Spherical structures 5.0-6.5 um in diam.

like those in the integumentary tissues are also present 76 within the nucellus and megagametophyte that surrounds the central vacuole (Fig. 80,95).

Abortive ovules — A small number of A^_ mirabilis cones have ovules with cellularized nucellus which is probably the result of internal nucellar proliferation

(Fig. 97,99). Some of these ovules appear to have mature seed integuments and are in the same size range as those mature cones described earlier (Stockey, 1975a). These cones also include the smaller one figured by Wieland

(1935, PI. 7, Fig. 3). The nucellar cells, which in these ovules measure 50-70 um in diam proliferated into the cavity left when the megagametophyte aborted or failed to develop (Fig. 97). The ovule in Figure 99 aborted after cellularized megagametophyte had started to form near the seed chalaza. No ovules have been found in which this tissue has entirely filled the seed cavity as Burlingame

(1914, PI. XXVI, Fig. 27,28) has reported in A;_ angusti- folia.

Many ovules in mature A^ mirabilis cones show no embryos, or internal nucellar proliferation but only shrunken nucellar and megagametophyte tissue. About

25-35% of the mature cones of A^ mirabilis contain seeds with embryos. Within this group, those cones having seeds . with embryos, approximately 75-85% of the seeds appear fertile. This is comparable with cones of Araucaria . 77 rulei or A^ columnaris on New Caledonia. Most seeds ex­ ternally appear to have fully developed tissues within, yet only about 35-40% ever germinate even if planting conditions are carefully controlled {M. Francois

Guinaudeau, personal communication). * • Megaspore membrane — In mature seeds the megaspore • membrane when present is usually crushed and discontinuous

(Stockey, 1975a). Thomson (1905b), Eames (1913) and Bur­ lingame (1915) all note that araucarian megaspore mem­ branes are thin at all stages of development. Berg (1950) reports a thin megaspore membrane in A^_ bidwillii which is usually visible only near the base (chalazal end) of the megagametophyte and is totally lacking near the apex.

If a megaspore membrane is present in the young ovules

(Fig. 79,80,90), it is indistinguishable from the shrunken nucellus and the tissues Which were developed inside.

Megagametophyte — The tissue toward the center of the tri-lobed nucellar region probably represents megagametophyte, some of which may still be in the free nuclear condition (Fig. 79,80,90,95). Burlingame (1914) , and more recently Hodcent (1967) discussed the formation of the megagametophyte in the genus Araucaria in which over 2,000 free nuclei have been reported as the normal condition before cell wall formation takes place. Polli­ nation takes place in living araucarians while the 78 megagametophyte is in the free nuclear stage of develop­ ment (Burlingame, 1914). Megagametophyte fills the nucellar cavity in mature ovules of A^ mirabilis and is composed of -isodiametric cells (50-70 um in diam) commonly filled with cell contents that probably repre­ sent starch grains (Stockey, 1975a). Cell walls in the free nuclear megagametophyte were probably laid down centripetally (Fig. 80,95).

Archegonia are borne singly and not in archegonial complexes in the upper portion of the megagametophyte

(Stockey, 1975a, Fig. 19). Sunken archegonia are charac­ teristic of the genus Araucaria (Seward and Ford, 1906).

Agathis also exhibits sunken archegonia but in some species these are scattered throughout the megagametophyte tissue

(Seward and Ford, 1906; Eames, 1913). Archegonia become sunken by an apical’proliferation of megagametophyte tis­ sue. Some of the cells in this region elongate (Fig. 98) as in A. 'bidwillii developing megagametophytes (Berg,

1950).

Embryos — In most cones of Araucaria mirabilis of the large size range containing seeds with embryos, the embryos appear to be in the dormant stage or what Schopf

(1943) calls the telo-stage period of development. Tis­ sues of’ the shoot apex, calyptroperiblem, hypocotyl, and apical meristem of the root are present (Stockey, 1975a). 79

» There are two cotyledons per embryo with 6-8 provascular strands in each cotyledon, each with an accompanying resin canal in the outer portion near the protoderm of the cotyledon. Some remains of highly coiled suspensor cells are present near the micropylar end of the ovule just to the base of the calyptroperiblem or "root cap" region

(Pig. 100).

Preservation of delicate embryo tissues is excellent and similar in many respects to that seen in P^ patagonica embryos in which there is one spherical structure 5.0-6.5 um in diam per each cell. These structures have been sug­ gested to have a number of possible origins in the living embryo: protein bodies, starch grains, other stored food,

lipid bodies, condensed cytoplasm, or possibly nuclear material (Stockey, 1975a). Cecich and Horner (1977), in a study on Pinus banksiana, report the changes with time

in the shoot apex of embryos during imbibition. After

48 hrs., protein bodies are reported to be breaking down, the nucleus is uniformly granular, lipid bodies and starch are abundant. Since the silica permineralization of A. mirabilis probably initially involved imbibition of water

and silicic acid by the cones, one would expect the embryos

to undergo at least some of the first 48 hour changes be­

fore becoming silicified. The storage products in general decrease rapidly in P. banksiana during imbibition (Cecich 80 and Horner, 1977). There is an increase in protein body vacuoles and heterochromatin up to 168 hours in this spe­ cies. Because of their constancy in size range, regular structure, and placement, one per each cell, there is the possibility that the spherical bodies in the fossil embryos represent nuclear material (Fig. 101).

The tissues of the telo-stage embryo of A^ mirabilis are closely comparable to those of living araucarian and other conifer embryos (Burlingame, 1915; Buchholz and

Old, 1933; Schopf, 1943; Allen, 1947; Owens and Molder,

1975). Embryos attain a length of up to 5 mm and a cross- sectional diameter of approximately 1 mm in the mature state.

Some embryos appear to be in the early telo-stage period in which total embryo expansion has not occurred even though all the meristems of the dormant embryo are present (Fig. 101}. As in the extant taxon Larix laricina

(Schopf, 1943) all the tissues except secretory elements are present within the fossil embryo (Fig. 101). These embryos, which measure 550 um in diam, are contained with­ in some of the intermediate-sized cone's (Fig. 101) .

Calyptroperiblem makes up the most extensive tissue of these embryos. 81

• DISCUSSION

Cone structure# vascularization# pith structure, system of resin canals# and structure of the cone-scale complex of Araucaria mirabilis are similar to living species of the genus Araucaria. Except for its much smaller overall size range, Araucaria mirabilis closely resembles the living A^ bidwillii from Southern Queens­ land. Araucaria bidwillii cones commonly reach a size of

30 cm when mature while A^ mirabilis cones are 10 cm in maximum length and diameter. Similarities between the

2 species include the separate origin of bract and ovuli­ ferous scale vascular supplies from the cone axis stele, winged cone-scales, a large "ligular sulcus" or space between the bract and scale (1/2 to 1/3 the length of the scale), a strongly vascularized ovuliferous scale tip

(ligule), and a vascular plexus supplying the ovule and ovuliferous scale. Seed integuments also-show a similar zig-zag pattern of cells in the sclerotesta. palder

(1953) placed A^. mirabilis in the section Bunya of the genus Araucaria that contains only A;_ bidwillii. The taxon Araucaria mirabilis would seem to fit in the section

Bunya (Wilde and Eames, 1952). This section was defined by Wilde and Eames as follows: 82

Genus: Araucaria de Jussieu 1789 Section: Bunya Wilde and Eames 1952

Leaves large, flat, spreading or slightly imbri­ cate. Male cones axillary; female cones subsessile or on short (to 2 cm long) peduncles. Ovulate cone-scales large, heavy, with woody wings; dehiscent, .the large "seed" shed from the scale at maturity; vascular supply of bract-scale unit double at source. Germination hypogeal. Cotyledons long-stalked in germination, retained within seed coats, the stalks fused into a hollow cylinder. Seedling fleshy, with a long subterranean dormant period. A. bidwillii Hook. A. mirabilos (Speg.) Windhausen emend. Calder

Information about the position or appearance of male cones in A^ mirabilis is lacking, as well as the position of ovulate cones with respect to the rest of the plant. Cone peduncles in A. mirabilis are short, reaching up to 1.1 cm long.

There still remains a question as to the mode of seed dispersal in A^ mirabilis. Araucaria bidwillii is the only species known in this genus to actually shed the seed as well as the scale at maturity. Cones of A. bidwillii commonly drop to the ground and shatter. The seeds are a good food source and are usually eaten by wallabies, cockatoos, and other small animals as well as the grain Calandra oryzae•(Silivicultural notes,

1917) because of the large amounts of stored starch in the megagametophyte tissue. The evidence fdr Aj_ mirabilis 83 .

seed dispersal is varied. Isolated cone axes have been

found (Fig. 87,88) in small numbers in the Cerro Cuadrado

forest but no isolated cone-scales or seeds have been found

in these deposits. Gothan (1950) first suggested that

A. mirabilis did not shed its scales because of the lack

of isolated fossil cone-scales. Calder (1953) believed

this lack of isolated scales was due to the season in which

the cones were buried. Only a few naked axes were left

from the previous year, some cones were immature, and others had not quite reached maturity. The mature embryos

in A^ mirabilis cones in the. larger size range, however, would appear to indicate that unlike Araucaria section

Eutacta, the cone-scales were not shed at this stage but may have remained intact until cone drop as in A^ bidwillii.

Specimens such as those of Fig. 85, 86 might also indi­ cate the retention of the scales at maturity since they were probably mechanically removed in some cases. The loose seeds that are seen in some of the more porous

specimens are actually nucule casts (nucellar casts) and are the result of imperfect preservation of the seed inte­ guments and not actually seeds which have separated from

the ovuliferous scale tissue itself. On the other hand,

the thin layer of ovuliferous scale tissue over the seed

integument could have been easily broken loose to release

the seeds. There remains the possibility also that small 84 animals may have served as agents, burying the seeds and thus effecting germination.

The presence of hypogeal germination in the

sections Bunya and Columbea of the genus Araucaria is rare among conifers. Araucaria mirabilis embryos are dicotyledonous and similar to those produced by members of

these 2 sections of the genus in vasculature and resin canal distribution. Seedlings probably exhibited hypogeal germination, and hypocotyl swelling as is indicated by the fossil seedings in Figures 21, 24, and 25.

The double resin canal system in the cortex of A. mirabilis cone axes is similar to that reported for

A. araucana (Thomson, 1913). The connection of the corti­ cal resin canals with the resin canal system in the cone- 4 scale complex occurs also in living araucarians between the pith and the leaves in young twigs. Jeffrey (1908, 1910) regarded this as a primitive feature in Prepinus twigs in which cortical resin canals also connect with those in the leaves. This is unlike the situation in Pinus stems,, e.g., in which the two systems never connect, those in the leaves ending blindly. The resin canal system in fossil twigs assignable to A^ mirabilis on the basis of internal anatomical features should also have a- connecting resin canal system. The position of the resin canals with­ in the ovuliferous scale and bract, one canal underlying . 85 each of 10-12 vascular traces probably also reflects their position in the leaves of this species. This is similar to the resin canal distribution in A^_ rulei cones and twigs (Seward and Ford, 1906). % ' Integuments of A^ mirabilis are. very similar in development to those of A^_ bidwillii. Three distinct layers present in the young ovules and the presence of free nuclear megagametophyte correspond to the stage des­ cribed by Wilde and Eames (1948) as that of ovules shortly after pollination. Before pollination Wilde and Eames

(1948) found no integumentary differentiation. The vascular supply to the ovule is also insufficiently dif­ ferentiated to accurately determine at this stage. The outer layer of integument is continuous with the ovulifer­ ous scale tissue through which numerous resin canals

f pass. The two large resin canals which extend along the lateral ridge of the seed remain the most conspicuous as the seed coat begins to enlarge. The intermediate stage (Fig. 81, 82) of seed integument differentiation shows these 2 canals clearly. In the mature seed df A. mirabilis (Stockey, 1975a) the resin canal system is vir­ tually obliterated due to integument expansion. Most, integument enlargement is in the downward direction as is true in ovules of A^ bidwillii (Wilde and Eames, 1948).

At maturity the sclerotesta makes up the bulk of the seed 86 integument, and shows the intricate 2ig zag pattern o£ sclereids found in integuments of A;_ bidwillii. The sacrotesta and endotesta are only very thin layers of tissue as is typical of most conifer seeds at this stage of development (Quisumbing, 1925). The endotesta is usually slightly thicker at the chalazal end of the seed where it is attached to the nucellus.

Vascularization of the seed by many vascular strands from an anastomosing plexus is another feature A^_ mirabilis shares with A^ bidwillii. Up■to 4 vascular bundles have been observed in seeds of A^_ mirabilis, although there are ■ • i probably a larger number of strands in any one seed which are lost due to the 1.5 mm saw cut or that are obscured by sclerotesta expansion.

The nucellus of A^ mirabilis ovules is similar to that in other species in the Araucariaceae. Agathis australis (Eames, 1913), Araucaria bidwillii (Wilde and

Eames, 1948), A^ angustifolia Burlingame (1914), A^ rulei, and A^ araucana (Seward and Ford, 1906) all exhibit a

"stipitate" nucellus (Eames, 1913) which is attached to the endotesta near the chalaza and exhibits a tenuously folded or invaginated apex near the micropylar end of the seed. In young mirabilis ovules the nucellus extends out of the micropyle of the seed and appears truncated as in young A. bidwillii ovules at the pollination stage (Wilde and Earnes, 1948). The nucellus is cellular at this

•stage (Fig. 80, 95) and soon shows evidence of drying, megagametophyte expansion within, and pollen tube damage.

The cone in Figure 66 that Wieland believed to be staminate exhibits the youngest nucellar developmental stage (Fig.

94) and probably represents the youngest•stage of A. mirabilis cone development available in the- Cerro Cuadrado collections. There is no evidence of an internal vacuole « within the nucellar tissue, only a solid tissue mass with intracellular inclusions similar to those seen in the embryos. Integuments of these ovules are poorly preserved.

Whether this cone and another (P13972, Wieland, 1935 PI.

6 Fig. B) that shows a similar structure were abortive or merely arrested in this growth stage due to fossili- zation is not known. These 2 cones are comparable with those of A^ bidwillii before the pollination stage and were in the beginning of their first year of growth prior to fossilization.

There are some cones which do reveal distinctly abortive ovules (Fig. 97, 99) in which there has been an internal proliferation of the nucellar tissue and that show no evidence of prior megagametophyte development.

It is certain that these ovules aborted during their first year of - development because of this tissue proliferation which also takes place in ovules of A. anqustifolia 88

(Burlingame, 1914). The ovule in Figure 99 aborted after megagametophyte formation possibly during the second year of growth when an embryo failed to develop. Seed integuments on such abortive ovules may continue to ex­ pand arid develop similar to those in fertilized seeds.

This condition is also seen in extant araucarian ovules in which seed viability cannot be determined by the size of the seed (M. Francois Guinaudeau, personal com­ munication) .

The wavy apex of the nucellus may be due in part to pollen tube action within the ovule (Fig. 92}. At maturity the nucellus is very thin and appears very similar to that of Agathis australis illustrated by Eames (1913; PI. 1,

Fig. 4), The pollen in Araucaria lands on the ovuliferous scale tip and the pollen tube grows along in grooves on the upper scale surface (Berg, 1950) to the extended nucellus (Thomson, 1905a, 1907). This type of pollination is present only in Araucaria, Agathis (Araucariaceae),

Saxegothaea (Podocarpaceae), and Tsuga (Pinaceae) (Wilde and Eames, 1948, Doyle and O'Leary, 1935 a, b).

Pollen tubes grow into the upper part of the nucellus usually in a plane perpendicular to the main axis of the gametophyte (Berg, 1950). Eames (1913) notes the destruc­ tive capacity of these pollen tubes which may enter all the tissues of the cones (ovuliferous scale, vascular 89 tissue, pith, and cortex of the cone axis) . The wavy apex of Araucaria mirabilis nucellar tissue may contain some of these tubes. The cortex of the cone axis which is not well-preserved in some specimens may have also under­ gone pollen tube action as could the pith. Most cones with well-preserved axes, however, show no evidence of pollen tube digestion and yet have embryos within the seeds.

The megagametophyte tissue of mirabilis appears to be in a couple of distinct developmental stages. The probable free nuclear stage illustrated in Figures 80 and 95, for example, suggests that the direction of wall formation was centripetal, the same as that in most extant conifers. Small lens-ahaped to spherical inclusions with­ in the v.acuole of the nucellus are well-preserved and may represent nuclear material.

At later developmental stages the megagametophyte tissue is composed of isodiametric parenchyma cells which appear to have granular cell contents as well as spherical cellular inclusions (6-7 um in diam) (Stockey, 1975a).

Cell walls are delicate and often not as well-preserved as cellular contents. Mo ovules have been found in which only cellularized megagametophyte is preserved without embryo formation. The structure of the cells in this tissue would indicate a Pinus type of megagametophyte formation (Chamberlain, 1935) as opposed to that seen in 90

Taxus e.g., in which elongate alveoli are laid down before

the addition of cell walls. Berg (1950) reports that megagametophyte development in A^_ bidwillii is from the

base upward, i.e., young gametophytes are located near the

chalazal end of the seed and then elongate acropetally.

This enlargement of the megagametophyte is mainly due to

cell elongation. Cells nearer the micropylar end of A. mirabilis too show elongation of some of the megagameto­

phyte cells (Pig. 98).

Berg (1950) reports 15-20 archegonia per ovule in

A. bidwillii although as few as 5 have been noted. Arche­

gonia are present in A^_ mirabilis and borne singly rather

than in complexes in the upper portion of the megagameto­

phyte. They are sunken in the megagametophyte tissue

(Stockey, 1975a) as are those in Agathis australis (Eames,

1913) which are superficial in origin and become sunken

with subsequent proliferation of the megagametophyte

cells which results in a channel leading to the neck

cells of the archegonium. This channel in A;_ bidwillii

may reach 400-800 um (Berg, 1950). No neck cells or

jacket cells have been Observed in A^ mirabilis.

The megaspore membrane is delicate and thin at

all stages of development in the genus Araucaria, and

is usually only apparent near the chalazal end of the ovule

(Berg, 1950), Thomson, 1905b; Burlingame, 1915). In Araucaria mirabilis traces of a megaspore membrane are visible in-some seeds (Stockey# 1975a) but for the most part agree with the condition seen in living species of the genus. In other conifer groups# the Pinaceae# Taxo- diaceae, and Cupressaceae, for example prominent megaspore membranes are known averaging 3-5 urn thick in Pinus,

Sciadopitys# and Thuja# and are thickest near the chalazal end of the ovule (Thomson, 1905b).

The dicotyledonous condition of A^ mirabilis is seen to have been present by the Jurassic period. Buch- holz's (1919) idea that polycotyledony is more primitive in' conifers will probably have to be studied in fossils from sediments older than the Jurassic Cerro Cuadrado material# since both types of embryo, dicotyledonous A^ mirabilis embryos# and the polycotyledonous embryos of Pararaucaria patagonica are both present during this period.

Stages of proembryo development have not been ob­ served in A^ mirabilis. This is probably due to a number of factors. Firstly, the forest was overwhelmed by vol- canics in a relatively short period of time as evidenced by the excellent state of preservation. Whether volcanics covered the vegetation at different times at the separate collecting localities is uncertain. It would seem, how- ever, that for the most part the vegetation was buried during one season. Ovules with free nuclear 92 megagametophyte and seeds with mature embryos would indi­ cate that cones of A^_ mirabilis like those of other conifer taxa took 2-3 years to develop to the stage of shedding mature seed. Rare instances among these fossils show * slightly earlier developmental stages than the two most frequent ones, e.g., the cellularized nucellus from the cone in Figure 66 (Fig. 94) without a developed mega­ gametophyte, and early telo-stage embryos (Fig'. 101) with no resin canals yet differentiated. These slight dif­ ferences can, however, also occur in extant araucarians, as well, during 1 season. Locality information in some collections is not specific, but there does seem to be a representative example from the Cerro Cuadrado (Field

Museum), the Cerro Alto or Cerro Chato (British Museum and Stockholm Riksmuseet), and the Cerro Madre e Hija

(Field Museum, Spegazzini's material (1925), and the specimens studied by Menendez (1960). Much of the material sold in London and Stockholm was purchased by Mansfeld at a number of the estancias nearby. Yet among the material • from these diverse•locations the 2 major developmental stages of A^ mirabilis cone development predominate sug­ gesting a large amount of volcanism that was probably very extensive geographically. It would seem possible, therefore, to predict the season in which the forest was inundated. Burlingame (1914, 1915) reports that the free

i 93

nuclear stage of megagametophyte development of A. angus-

tifolia occurs in January-February. and that seed drop in

cones from the previous year occurs between November and

January. Thus, at the time of pollination of one crop

of cones the seeds are being shed from another. Although

these times vary somewhat year to year, it would appear

the A^ mirabilis cones with intact seeds and cone-scales

should have shed their mature seeds (indicated by the -

presence of a free nuclear stage in young cones) if their

life cycle coincided closely with that of 2W angustifolia.

Wilde and Eames (1948) and R. J. Gould (personal communi­

cation) report a longer life cycle for A;_ bidwillii. In

A. bidwillii, grown in Southern Queensland where it is

native, the free nuclear stage of megagametophyte develop­

ment takes place from August to October and seeds are shed

from January to March. Therefore, on any one during

the period of October-December one might expect to find

cones of comparable developmental stages to those seen

in A^_ mirabilis at the Cerro Cuadrado.

In A;_ araucana (Montaldo, 1971), the closest species

to the Cerro Cuadrado geographically, young cones are1 in

the free nuclear stage until December and seeds are

mature and shed by March or April. In its native Andean

■ habitat dispersed seeds remain beneath a snow cover from

April to November and seedlings germinate when the snow 94 melts in November (Montaldo, 1971). This life cycle inr formation would closely parallel the situation seen in

A. mirabilis in which free nuclear megagametophyte stages and mature embryos are present, and would also explain the young seedling structures observed in the forest (Fig. 21, 24, 25) which represent first year seedling growth stages.

The mature or telo-stage embryo of A_^ mirabilis is very similar to other conifer embryos at a comparable developmental stage. Calyptroperiblem ( root cap ), apical meristem of the root, hypocotyl, shoot apex and cotyledons present in all embryos seen in A^ mirabilis indicate a relatively mature embryo. One cone with, slightly smaller embryos and cotyledons without the ring of resin canals beneath the epidermis probably indi­ cates an early telo-stage embryo. Schopf (1943) reports a similar developmental sequence in Larix. Unlike the dormant embryo of Cedrus (Buchholz and Old, 1933) however,

A. mirabilis embryos do not show evidence of the first leaf primordia at this developmental stage.

The excellent preservation of the Cerro Cuadrado material has enabled the elucidation of much of the life cycle of the fossil conifer Araucaria mirabilis. Im­ mature developmental stages of cones which are necessary for the completion of developmental studies have not previously been recorded, and provide the only basis on which conifer phylogeny may be based. It is hoped that some of the structural information presented here may be used in less well preserved conifer material, enabling the recognition of developmental stages of one organism, and to provide a basis for understanding the evolution of some of the more detailed parameters of conifer phylogeny. GENERAL DISCUSSION

The main interest in the Cerro Cuadrado fossil conifers historically has been their geographic distri­ bution and their relationship to other conifer groups.

While these parameters are important, a greater amount of information about their morphology, ontpgeny and repro­ ductive biology has become available through the use of more sophisticated techniques. The unusual state of preservation of this material including delicate seed and embryo tissues as well as cellular contents has not only allowed for close comparisons with living conifers but has also provided us with information about the more subtle biological phenomena such as seedling germination and development, various embryological stages, seed inte­ gument differentiation, megagametophyte development, seed dispersal and other reproductive mechanisms.

The ovulate cone Pararaucaria patagonica combines an interesting set of characters which are common to four or five families of living conifers. The one-seeded condition of araucarians is one of the features linking

Pararaucaria to the Araucariaceae. The generic name

Pararaucaria is a misnomer indicating a relationship that probably does not exist. 96 97

Vascularization of the cone-scale complex is

similar to that in the Pinaceae. Miller (1976) has sum­ marized the information known to date on the vascu­

larization of fossil and living members of the Pinaceae.

Although the vascularization of Pararaucaria does not

closely resemble any of the fossil or living pinaceous

genera, it is most similar to that observed in Pseudo­

araucaria Fliche emend. Alvin (1957a) , with an abaxially

concave scale trace and a single bract trace which are

separate at their origins. The lack of resin canals

in the xylem of the cone axis of Pararaucaria is

common in the Taxodiaceae as well as the Araucariaceae.

Pith structure is similar to Taiwania cryptomerioides

Hayata (Doyle and Doyle, 1948) in the Taxodiaceae.

The multi-layered seed integuments of Pararaucaria

are similar to those in the Cupressaceae and Taxodiaceae

and are most like those described by Konar and Banerjee

(1963) in Cupressus funebris. Seed win^s probably arose

from the upper surface of the ovuliferous scale similar

to those in members of the Pinaceae.

Pararaucaria was a conifer with taxodiaceous and

pinaceous affinities which is different enough in its

combination of characters to warrant placement in its own

family. It has polycotyledonous embryos which probably

exhibited epigeal germination and photosynthetic cotyledons. 98 Its winged seeds were dispersed and show an abscission layer at their points of attachment. Cones may have been closed after pollination by resinous substances produced by the glandular trichomes or hairs on the upper and lower ovuliferous scale surfaces as occurs in living

Cedrus (Chowdhury, 1961}.

Araucaria mirabilis cones were probably borne on branches with rhomboidal leaf scars and imbricate ever­ green foliage. The seedlings which were dicotyledonous, probably exhibited hypogeal germination ancl pronounced hypocotyl swellings.. At the time of pollination the integuments were three layered and the nucellus extended out of the micropyle as in living araucarians. Megaga­ metophyte development was similar to that in A. bidwillii

(Berg, 1950), A^ angustifolia (Burlingame, 1914), and

A. araucana (Hodcent, 1967) and probably involved acropetal elongation after cell wall formation. Wall formation in the megagametophyte is the Pinus-type

(Chamberlain, 1935) and results in isodiametric cells; cell wall formation beginning in the periphery. Arche-

* gonia were sunken in the upper portion of the megagameto­ phyte and borne singly.

It is possible that unlike most living araucarians,

A. mirabilis cones shed their seeds instead of their scales at maturity. Although two specimens that represent 99 naked axes have been found (Fig. 87, 88), most cones with mature seeds are intact and those with missing cone-scales seem to have been mechanically removed (Fig. 85). Cones may have shattered, shedding both seeds and scales on impact as in A^ bidwillii from southern Queensland.

Araucaria mirabilis cones probably represent those of an extinct species of the Araucariaceae and show the most similarity to Araucaria bidwillii with respect to cone structure, vascularization, strongly vascularized ovuliferous scale tip, sclerotesta structure, and the vascular plexus supplying the ovule and ovuliferous scale.

The author agrees with Calder (1953) that A^_ mirabilis cones meet the criteria listed by Wilde and Eames (1952) for inclusion within the section Bunya of the genus

Araucaria.

There is one part of the life cycle of these fossil conifers which is conspicuously absent, the pollen cones, including the microgametophyte phase of the reproductive cycle. There have been only three microsporangiate cones found at the Cerro Cuadrado Petrified Forest.

These were relatively imperfectly preserved and given the organ.genus name Masculostrobus (Menendez, 1960). The preservation of Masculostrobus altoensis Menendez was not such that it could be placed in any conifer family. Per­ haps most pollen cones have been overlooked, and like 100 young cones of L mirabilis, they are embedded'in the volcanic ash matrix. Further collecting might fill this gap in our knowledge of the life cycle.

No pollen grains, pollen tubes or what could be interpreted as pollen tube damage have been previously reported in A^ mirabilis or patagonica cones. Arau- carian pollen tubes are well-known for their destructive capacity in ovule tissues during growth. (Thomson, * 1905a, 1907). They invade all the cone tissues and have been known to grow through the pith and even vascular tissues of the cone axes. Some cones show imperfectly preserved pith and cortical cells in their axes. Tears or holes in this material may be the result of pollen tube action. The wavy crushed cells near the apex of the nucellus may be partially a result of pollen tube action. The vascular tissues of the cone axes are seldom interrupted by holes or tears unless the entire cone is preserved in such a way. These cones with extensive deterioration of tissues may show the results of fungal action. Most cones of A. mirabilis examined during this study, however, show no visible evidence of pollen tube digestion and yet have embryos within the seeds.

Seedlings representing an additional segment of the life cycle of the Cerro Cuadrado conifers have been studied anatomically. The two original "seedling" 101 specimens described by Wieland (1935) probably belonged to Pararaucaria patagonica. Wood is of a similar type to that of the cone axes, and the pith structure is also simi­ lar resembling Taiwania cryptomerioides Hayata (Doyle and Doyle, 1948).

Most of the tuberrlike or corm-like structures, however, are probably seedlings that were produced by

Araucaria mirabilis. Their similarity in structure, to first year seedlings of the genus Araucaria in sections

Bunya and Columbea is striking. They exhibited hypogeal germination and a pronounced hypocotyl swelling, a rare seedling type among conifers.

Reports of seedlings in the fossil record are rare.

The only fossil conifer seedlings known to the author are those of Araucarites phillipsi from the Jurassic of

Yorkshire (Kendall, 1949) and Aethophyllum stipulare

Brongniart from the Triassic of France (Grauvogel-Stamm and Grauvogel, 1975) once thought to be a member of the

Eguisetales, because of a lack of reproductive organs, but reinterpreted as a conifer species. Both are inter­ preted as having epigeal germination. The cone-scale of Araucarites phillipsi described by Kendall is illustra­ ted as germinating (Fig. J, 1949) from a winged Eutacta- type cone-scale. This was a compression specimen, however, and no internal anatomical details are preserved. 102

The large numbers of specimens and excellent preser­ vation of the Cerro Cuadrado materials has provided the type of information necessary in the interpretation of conifer phylogeny. Immature developmental stages of conest and young sporophytes which are necessary to com­ plete life cycles of these fossil conifers here provide a structural basis which may be used in less well pre­ served conifer material for interpretations of different developmental stages. 4

APPENDIX

ti

Figure Explanations

Fig. 1-7. Pararaucaria patagonica. b=bract, bt=bract

trace, c=cortex, os^ovuliferous scale, ostgovuliferous scale trace, p=pith, s=seed, sc-sclerenchyma, w=winq, x=xylem. 1. External surface Of ovulate cone. 031218.

X 2. 2. External surface of less weathered cone. Arrow

indicates striation of ovuliferous scale external surface.1

P13988. X 2. 3. Outermost portions of the cone-scale complex showing large cells with, amorphous contents.

P139— I #1. X 25. 4. Longitudinal section of the cone

axis at a level of trace divergence. P13978 A #1. X 7.

5. Transverse section of cone axis. P13961 #2. X 14.

6. Longitudinal section of cone-scale complex showing position of seed. P13978 A #2. X 10. 7. Radial longi­

tudinal section of cone axis with unseriate circular-

bordered pits of tracheids. P139— I #1. X 250.

103

105

Fig. 8-13. Seed and cone-scale features. bt«bract trace, end=endotesta,, os t=*o vu 1 iferous scale trace, s«seed, sar= safcotesta, sc=sclerenchymaf scl=sclerotesta,'vb«vascular bundle, w=wing. 8^ Pararaucaria patagonica seed integu­ ment. P13979 #4. X 36. 9. P. patagonica seed wing.

Arrow indicates trichome tip. Note branching of individual trichomes. P13958 #1. X 170. 10. P^ patagonica trans­ verse section of two-seeded cone-scale complex. P13979 side #3. X 12. 11. P^_ patagonica tangential section of cone showing vasculature of cone-scale complex. P139--

III A #2. X 22. 12. Abies grandis seed wing surface.

Arrows indicate bulbous tips of cells. X 200. 13. P. patagonica. Arrow indicates seed abscission layer.

P13978 B #1. X 12.5. 106 107

Fig. 14-20. Pararaucaria patacronica embryos. cal= calyptroperiblem, cot=cotyledon, h=hypocotyl, jainte- gument, m-megagametophyte, n=nucellus, ostcovuliferous scale trace, rm=root meristem, sa=shoot apex, sc=scleren- chyma, w=wing. 14. Longitudinal section of telo-stage embryo. 031217 #1. X 28. 15. Oblique embryo section with cellular megagametophyte. 031217 #1. X 30. 16.

Embryo tissue with spherical cellular inclusions. 031218

#2. X 300. 17. Transverse section of embryo with 8 cotyledons. 031218 #4. X 20. 18. Transverse section of embryo at cotyledonary node. 031218 #2. X 22. 19.

Transverse section-of embryo in hypocotyl region. 031217

#2. X 22. 20. Transverse section of embryo in calyp- troperiblem or "root cap" region. 031218 #2. X 25.

109

Fig. 21-27. Seedlings. P=fiith/ sx=secondary xylem.

21. Araucaria mirabilis seedling. BM1077. XI. 22.

Transverse section of seedling in Fig. 24 at level A.

BM1078 #5. .X 3.3. 23. Transverse section of seedling in Fig. 24 at level A. Resin canals indicated by arrows.

BML078 #5. X 8.3. 24. Araucaria mirabilis seedling.

BM1078., X 1. Araucaria mirabilis seedling showing tru- cated shoot axis. BM1081. XI. 26. Transverse section of seedling in Fig. 24 at level A, showing large pith cells and position of secondary xylem. BM1078 #5. X 27. 27.

Transverse section of seedling in Fig. 24 at level A showing isodiametric cortical cells. BM1078 #4. X 30.

,11 1

Pig. 28-33. Seedlings. ca=cambium. cb=cotVledonarv bundle. cotgcotvledon. cs=cone-scale. h=hvpocotyl. m= «: megagametophyte. pb~plumular bundles, pe=periderm, pp= primary phloem, sp=secondary phloem, sx=secondary xylem.

28. Germination sequence of Araucaria heterophylla from

* emergence of radicle from the cone-scale complex to cotyledon emergence, hypocotyl extension, cotyledon sepa­ ration, and branch initiation (arrow). X 0.4. 29.

Araucaria angustifolia seedling showing swollen hypocotyl and cotyledons embedded in megagametophyte tissue. Seed coats have been removed. X 0.9.■ 30. Transverse section of seedling hypocotyl vascular bundle showing primary phloem, secondary phloem, cambium, secondary xylem, and primary xylem points (arrows). D bot #5. X 92.

31. Araucaria bidwillii seedling with swollen hypocotyl after 4 months XI. 32. Araucaria angustifolia seedling with elongate primary.root, swollen hypocotyl and newly initiated shoot (arrow). X 0.9. 33. Transverse section

Of A^ bidwillii seedling in Fig. 36 at level A, with an external periderm, fused plumular bundles, cotyledonary bundles, and resin canals (arrows). A^ bid. #3 E top #33.

X 9. S a s « ^ ^ B ||| m i W 3^M4^£9h b E issst

;*>>*> \< 113

Pig. 34-37. Seedlings. cot=cotyledon, pe=periderm.

34. Transverse section of Araucaria bidwillii seedling

in Fig. 36 at level B showing 5-parted symmetry, massive cortex with associated resin canals, separate vascular

bundles, and periderm. bid. #3 E top #49. X 7. 35.

Pararaucaria* patagonica seedling figured by Wieland.

Arrows indicate the position of lateral roots. Patches of periderm are still attached, to the surface. P14614A.

X 1. 36. Araucaria bidwillii seedling with cotyledons,

swollen hypocotyl, primary and lateral roots. X 0.8.

37. Transverse section of swollen hypocotyl of A/ angus­

tifolia seedling in Fig. 29. Periderm arises just outside

the outer ring of resin canals. Resin canals occur in a

concentric ring around the 4 vascular bundles and scat­

tered among the starch filled cortical cells. D bot #5.

X 16. 114 115

Pig. 38-44. Seedlings. ct=cotyledonary tube, i=inte-

gument. 38. Germinating seedling of Araucaria bidwillii.

Cotyledonary tube is still attached to the seedling that- has emerged from the integuments. Arrow indicates the positioh of the primary root. XI. 39. Pararaucaria patagonica seedling illustrated by Wieland. Arrows in­ dicate the position of lateral roots. P14614 B. X I . 40.

View of fossil seedling showing axis (arrow). SR2024.

X 1. 41.' Tangential section of seedling wood in Fig.

35, showing uniseriate ray in the secondary xylem. P14614

A side #1. X 360. 42. Transverse section of seedling

in Fig. 35, showing decorticated stem containing several growth rings in the secondary xylem, and uniseriate rays

lacking resin canals. P14614 A. #1. X 9. 43. Underside of seedling in Fig. 40. SR2024. X I . 44. Lateral view of seedling in Fig. 40. SR2024. X 1. W M m m

m m 117

Fig. 45-53. Seedlings. 45. Corm-like fossil seedling.

SR2025. X 1. 46. Fossil seedling in Fig. 45 undersur­

face. SR2025-. X I . 47. Corm-like fossil seedling.

BM1071. X 1. 48. Fossil seedling in Fig. 47 undersur­

face. BM1071. X 1. 49. Fossil seedling in Fig. 47

lateral view. BM1071. X I . 50. Corm-like fossil

seedling. BM1072. X 1. 51. Fossil seedling in

Fig. 50 undersurface. BM1072. X I . 52. Fossil

seedling in Fig. 50 lateral vi6w. BM1072. XI. 53.

Transverse section of decorticated fossil seedling in

Fig. 40. SR2024. #5. X 4. 118 1 1 9 Pig. 54-57. Seedlings. a=axis, sx=»secondary xylem, t=tracheids. 54. Decorticated shoot-like structure.

BM1076. XI. 55. Tangential section of partially decorticated seedling in Fig. 56, showing expanded region

* with secondary xylem tracheids present in the outer and inner zones. Arrows indicate vascular rays,. A lateral branch is present with abundant secondary xylem con­ taining possible cambium and phloem. BM1080. side #3.

X 4. 56. Partially decorticated seedling showing a lateral branch and main axis. BM1080. XI. 57.

Decorticated shoot-like structure. BM1075. X 1.

121

Fig. 58-68. Araucaria roirabilis. a=axis, b=bract, os-ovuliferous scale, s=seed. 58. Immature cone with attached peduncle; arrows indicate laminar tips of the bracts. SR2014. XI. 59. Immature cone, view of pedun­ cle. SR2017. X I . 60. Partially weathered immature cone. SR2012. X I . 61. Tangential section of im­ mature cone in volcanic ash matrix. Arrows indicate nucellar tissue. V. 31414. X 2. 62. Immature cone.

SR2017. XI. 63. Weathered immature cone with exposed seeds. SR2019. X I . 64. Immature cone with attached peduncle. SR2016. XI. 65. Immature cone in volcanic ash matrix. V. 31414. X 2. 66. Immature cone.

P13972. X 1. 67. Intermediate sized cone. P13888.

XI. 68. Intermediate sized cone. Weathered surface shows exposed seeds. P13828 I. X I . 122 123

Pig. 69-76. Araucaria mirabilis. b=bract, bt=bract trace, os=ovuliferou3 scald, r=resin canal, s«seed. 69.

Intermediate sized cone with peduncle and weathered surface.

SR2018. X 1. 70.- External surface of weathered in­ termediate sized cone. P13935. X I . 71. Transverse section of cone in peduncle region. Arrows indicate po­ sition, of cortical resin canals. .P13828 II. #4. X 7. 72.

Intermediate sized cone with attached peduncle. 031221

E. XI. 73. Tangential section of cone near axis showing traces to the cone-scale complex and system of resin ca­ nals. Arrows indicate the position of ovuliferous scale trace departure. SR2016 #5. X 12. 74.' Radial longi­ tudinal section of secondary xylem of cone axis. P13939

#4. X 350. 75. Resin canal epithelium in bract apophysis.

P13929 #1.' X 160. 76. Tangential section of immature cone showing resin canal system in cone-scale complex. P13973

B #2. X 10.

I

125

Pig. 77-82. Araucaria mirabilis. b°bract, os=ovuliferous scale, r=resin canal, sc=sclerenchyma. 77. Longitudinal section of resin canal in bract. Arrows indicate epithelial lining. P13939 #1. X 26. 78. Transverse section of cone showing resin canals (arrows) and overlying vascular bundles in bract apophyses. P13929 #1. X 10.

79. Tangential section of young cone with immature ovules.

SR2016 #3.. X 22. 80. Young ovule with three-layered integument, ovuliferous scale resin canals, and cellula- rized nucellus. SR2016 #3. X 62. 81. Transverse section of intermediate stage cone-scale complex. Lateral ridge of the ovule is represented by two prominent resin canals.

P13935 #1. X 15. 82. Tangential section of intermediate stage cone showing ovules with prominent lateral ridges.

P13935 #1. X.5.

127

Pig. 83-90. Araucaria mirabilis. b=bract, ls=ligular sulcus, r«resin canal. 83. Paradermal section o£ bract apophysis showing 12 resin canals (some with contents). .

BM1067 #3. X 10. 84. Outer surface of cone showing weathered resin canal scars. P13935. X 2.1. 85. Cone piece with peduncle, and exposed axis. Cone-scale com­ plexes were mechanically removed. Arrow indicates half'of cone-scale complex. P13914. X I . 86. Cone with axis region removed. P13836. X 1. 87. Cone axis with rhomboidal scars and attached peduncle. 'P13981. XI.

88. Cone axis pieces. P13973. .XI. 89. Transverse ■ section of specimens in Pig. 88, showing A^ mirabilis pith, ring of vascular bundles (arrows), and inner and outer resin canal systems. P13973 A #1. X 9. 90.

Longitudinal section of cone-scale complex showing wide ligular sulcus and attachment of nucellar tissue to seed integument (arrow). SR2016 #5. X 8.

* 129 Fig. 91-96.. Araucaria mirabilis. a=axis, b=bract,

i=intequment, m=meqaqametophvte, n=nucellus, os=ovuli-

ferous scale, r«resin canal, s=seed, sx=secondary xylem.

91. Longitudinal section of cone with seeds embedded in ovuliferous scale tissue. Note absence of ovules in lower

portion of cone. P13939 #1. X 5. 92. Two seeds from a mature cone with wavy nucellar tissue near apex. Ar­

row indicates the position of the micropyle. P13939 #1.

X 9. 93. Vascularization at seed chalaza (arrows).

P13892 #5. X 15. 94. Cellularized nucellar tissue with

inclusions (arrows) extending out of seed micropyle.

P13972 B #1. X 170. 95. Celluarized nucellus and proba­

ble free nuclear megagametophyte with spherical 6 urn in­

clusions. SR2016 #3. X 162. 96. Vascular plexus

(arrows) at seed chalaza. P13892 #7. X.13.

131

Fig. 97-101. Araucaria mirabilis. cot^cotyledon. i=inte- qument, m=megaqametophvte, n=nucellus, su=su3pensor. 97.

First year aborted ovule with internal nucellar prolifera­ tion. P13820 I A.#1. X 10. 98. Apical portion of the megagametophyte from a mature seed showing some elongate cells. P13865#2. X 100. 99. Second year aborted ovule with celluarized magagametophyte near the base and inter­ nal nucellar proliferation. P13865 #2. X 10. 100.

Longitudinal section of seed showing suspensor remains in micropylar region. P13892 #3. X 60.2. 101. Transverse section of young telo-stage embryo with surrounding megagametophyte. Cotyledons each having 8 vascular bundles. SR2018 #3. X 95. 132 LIST OF REFERENCES

Aase, H. C. 1915. Vascular anatomy of the megasporo- phylls of conifers. Bot. Gaz. 60: 277-313.

Allen, G. S. 1947. Embryogeny and the development of the apical meristems of Pseudotsuga. III. Develop­ ment of the apical meristems. Amer. J. Bot. 34: 204-211.

Allen, T. C., 1951. Deformities caused by insects, p. 411-415. In F. Skoog

Alvin, K. L. 1953. Three abietaceous cones from the Wealden of Belgium. Mem. Inst. R. Sci. Nat. Belg. 125: 1-42.

_ . 1957a. On Pseudoaraucaria Fliche emend., a genus of fossil pinaceous cones. Ann. Bot. N.S. 21: 33-51.

_ . 1957b. On the two cones Pseudoaraucaria heeri (Coemans) nov. comb, and Pityostrobus villerotensis nov. sp. from the Wealden of Belgium. Mem. Inst. R. Sci. Nat. Belg. 135: 1-27.

Archangelsky, S. 1968. On the genus Tomaxellia (Coni- fferae) from the Lower Cretaceous of Patagonia (Argentina) and its male and female cones. J. Linn.' Soc. (Bot.) 61: 153-165.

Arnoldi, W. 1900. Beitrage zur Morphologic einiger Gymnospermen. V. Weitere Untersuchungen der Embryogenie in der Families der Sequoiaceen. Bull. Soc. imp. nat. Moscow. N.S. 14: 449-476.

Bartholomew, R. L., L. C. Matten, and E. F. Wheeler. 1970. Staining silicified woods. J. Paleont. 44: 905-907.

133 134

Beck* E. G., 1954. The nature of the stimulus in the Solldaqo gall induced by the larve of Gnorimosfchema gallaesoliaginis. Brookhaven Symposia in Biology No. 6. Abnormal and Pathological Plant Growth. Brookhaven National Laboratory, Upton, New York, p. 235-251.

Berg, D. H. 1950. The morphology of the gametophytes, fertilization and proembryo of Araucaria bidwilli (Hook.). Ph.D. dissert. Cornell University. 62 p.

i Bierhorst, D. W. 1971. Morphology of vascular plants. The Macmillan Company, New York.

* Blanchard, J. 1892. Curieuse germination de ,11Araucaria bidwillii. Rev. Hort. (Paris) Prance, p. 295.

Buchholz, J, T. 1919. Studies concerning the evolutionary status of polycotyledony. Amer. J. Bot. 6: 106- 119'.

______; 1939. The morphology and embryogeny of Sequoia gigantea. Amer.. J. Bot. 26: 93-101.

______. and E. M. Old. 1933. The anatomy of the embryo of Cedrus in the dormant stage. Amer. J. Bot. 20: 35-531

Burlingame, L. L. 1913. The morphology of Araucaria brasiliensis. I. ‘ The.staminate cone and male gametophyte. Bot. Gaz. 55: 97-114.

______. 1914. The morphology of Araucaria brasiliensis. II. The ovulate cone and female gametophyte. Bot. Gaz. 57: 490-508.

1915. The morphology of Araucaria brasiliensis. III. Fertilization, the embryo and.the seed. Bot. Gaz,. 59: 1-38.

Butts, D., and J. T. Buchholz. 1940. Cotyledon numbers in conifers. 111. Acad. Sci. Trans. 33: ,58-62.

Calder, M. G. 1953. A coniferous petrified forest in Patagonia. Bull. Brit. Mus. (Nat. Hist.) Geology 2: 99-138. 135

Cecich, R. A. and H. T. Horner, Jr. 1977. An ultra- structural and microspectrophotometric study of the shoot apex during the initiation of the first leaf in germinating Pinus banksiana. Amer. J. ' Bot. 64: 207-222. ' ”

Chamberlain, C. J. 1935. Gymnosperms: Structure and evolution. Univ. Chicago Press. « Chowdhury, C. R. 1961. The morphology and embryology of Cedrus deodara {Roxb.) Loud. Phytomorphology 11: 283-304.

Darrow, B. S. 1936. A fossil araucarian embryo from the Cerro Cuadrado of Patagonia. Bot. Gaz. 98: 328-337.

Darwin, C. . 1839. Journal of researches into the geology and natural history of the various countries visited by the H.M.S. Beagle. Henry Colburn. London.

Doyle, M. H. and J. Doyle. 1948. Pith structure in conifers. 1. Taxodiaceae. Roy. Irish Acad., Dublin, Proc. 52: 15-39;

Doyle, J. and M. O'Leary. 1935a. Pollination in Saxegothaea. Royal Dublin Soc. Sci. Proc. '(N.S.) 21: 175-179.

______, and . 1935b. Pollination in Tsuga, Cedrus, Pseudotsuga, and Larix. Royal Dublin Soc. Sci. Proc. (N.S.) 21: 191-204.

Durer, F. J. 1865. Ueber das Keimung und die Vermehrung der Araucaria bidwillii. Gartenflora 14: 103-105.

Eames, A. J. 1913. The morphology of Agathis australis. Ann. Bot. 27: 191-204.

Favre-Duchartre, M. 1963. A propos des nucelles des plantes a ovules. Ann. Arers. Tome 1, Fasc. 2: 35-42.

Feder, N.and'T. P. O'Brien. 1968. Plant microtechnique: Some principles and new methods. Amer. J. Bot. 55:

Feruglio, E. 1949. Descripcion Geologica de la Patagonia 1. Buenos Aires. 136

_ . 1951. Piante del Mesozoico della Patagonia. Pubbl. 1st. geol. Univ. Torino,, 1: 35-80.

/ Fliche, P. 1895. Etudes sur la flore fossile de l'Argonne (Albien-Cenomanien). Bull. Soc. Sci. Nancy. 14: 114-306.

Florin, R. 1940. The Tertiary Fossil conifers of South and their phytogeographical significance, with . a review of the fossil conifers of southern lands. . Kgl. Svensk. Vetensk. Acad. Handl. Stockholm. Ill, 19: 1-107. 4 _ . 1944. Die Koniferen des Oberkarbons und des Unteren Perms. Heft 7. Paleontographica B. 85: 457-654.

Fossa-Mancini, E. 1941. Los "Bosques Petrificados de la Argentina," segun E. S. Riggs y G. R. Wieland. Notas Hus. La Plata. 6 (Geol. 12): 59-92.

Frenguelli, J. 1933. Situacion estratigrafica y edad de la "Zona con " al sur del curso inferior del Rio Deseado. Bol.inf. Petroler. 112:.843-900.

Gordon, W. T. 1936. Petrified araucarian cones from Patagonia. Proc. Geol. Soc. Lond. 1317: 14-15.

Gothan, W. 1925. Sobre Restos de Plantas Fosiles pro- cedentes de la Patagonia. Bol. Acad. Nac. Cienc. Cordoba 28: 197-212. I' ______. 1950., Uber die merkwurdigen feigenartigen Kieselknollchen aus dem versteinerten Wald des Cerro Cuadrado in Patagonia. Misc. Acad. Berol. 1: 149- 154.

Grauvogel-Stamm, L. et L. Grauvogel. 1975. Aethophyllum Brongniar-t, 1828, Conifere (non Equisetale) du Gres S Voltzia (Buntsandstein Superieur) des Vosges (France). Note preliminaire. Geobios 8: 143-146.

Harris, T. M. 1943., The fossil conifer Elatides william- soni. Ann. Bot. N. S. 7: 326-339.

Heckel, E. 1892. Sur la germination des graines d 1 Araucaria bidwillii Hook. Bot. Cent. 52: 405-406. 137

Hemsley, W. B. 1901. Germinating seeds of Araucaria bidwillii. Proc. Linn. Soc. Lond. 1901*02. 114s l ^ T .

Hill, T. G. and E. de Fraine, 1908. On the seedling structure of gymnosperms * I. Ann. Bot. 22: 689* 712.

______, and ______. 1909a. On the seedling structure of gymnosperms. 1 II. Ann. Bot. 23: 189*227.

______, and ' 1909b. On the seedling structure of gymnosperms III. Ann. Bot. 23: 433-458.

Hirmer, M. 1936. Die Bluten der Coniferen. I. Entwick lungsgeschichte und vergleichende Morphologie des weiblichen Bliitenzapfens der Coniferen. Biblio. Bot. 23: 1-100.

______, and L. Horhammer. 1934. Zur weiteren Kenntniss von Cheirolepis Schimper und Hirmeriella Horhammer mit Bemerkeungen uber deren systematische Stellung. Paleontographica B. 79: 67-84.

Hodcent, E. 19,67. Etude cytobiologigue du prothalle femelle de 1'Araucaria araucana aux derniers stages cenocytiques et lors de son passage a l'etat "alveolaire". C.R: Acad. Sc., Paris 264: 2746- 2749.

Jeffrey, E. c. 1908. On the structure of the leaf in Cretaceous pines. Ann. Bot. 22: 207-220.

■ 1910. A new Prepinus from Martha's Vineyard; Proc. Boston Soc. Nat. Hist. 34: 333-338.

Johansen, D. A. 1940. Plant Microtechnique. McGraw- Hill Book Co., Inc. N. Y.

Jung, W. 1968. Hirmerella munsteri (Schenk) Jung nov. comb., eine bedeutsame Konifere des Mesozoikums. Palaeontographica. B. 122: 55-93.

Kendall, M. W. 1949. A Jurassic member of the Araucaria- ceae. Ann. Bot. N. S. 13: 151-161. 138

Konar, R. N. arid S. K. Banerjee. 1963. The morphology and embryology of CupreBsus funebris Endl. Phytomorphology. 13: 321-337.

' Lorch, J. 1967. A Jurassic flora of Makhtesh Ramon, Israel. Israel J. Bot. 16: 131-155.

Menendez, C. A. 1960. Cono masculino de una conifera fosil del Bosque Petrificado de Santa Cruz. Ameghiniana. 2: 11-17.

Miller, c. N. 1969. Pinus avonensis a new species of petrified cones from the Oligocene of Montana. Amer. J. Bot. 56: 972-978.

______. 1970. / Picea diettertiana, a new species of petrified cones from the Oligocene of Montana. Amer. J. Bot. 57: 579-585,

______. 1972. Pityostrobus palmeri, a new species of petrified conifer cones from the Late Cretaceous of New Jersey. Amer. J. Bot. 59: 352-358.

______. 1973. Silicified cones and vegetative remains of Pinus from the Eocene of British Columbia. Con- trib. Univ. Mich. Mus. Paleontol’. 24: 101-118.

______. 1974a. Pityostrobus hallii, a new species of structurally preserved conifer cones from the Late Cretaceous of Maryland. Amer. 'J. Bot. 61: 798-804.

. 1974b. Pinus wolfei, a new petrified cone from the Eocene of Washington. Amer. J. Bot. 61: 772-777.

. 1975. Petrified cones and needle bearing twigs of new taxodiaceous conifer from the Early Cretaceous of California. Amer. J. Bot. 62: 706- 713.

______. 1976. Early evolution in the-Pinaceae. Rev. Palaeobot. Palynol. 21: 101-117.

______, and C. R. Robison. 1975. Two new species of structurally preserved pinaceous cones from the Late Cretaceous of Martha's Vineyard Island, Massachu­ setts. J. Paleont. 49: 138-150. 139

Mitra, A. K. 1927. On the occurrence of two ovules on araucarian cone-scales. Ann. Bot. 41: 461-471.

Montaldo, P. R. 1971. Life history of Araucaria araucana (Mol-) Koch. M.Sc. Thesis. Washington State Universi- ty.

Owens, j. N. and M. Molder. 1975. Pollination, female gametophyte, and embryo and seed development in yellow cedar (Chamaecyparis nootkatensis). Can. J. Bot. 53: 186-199.

______, and F. H. Smith. 1965. Development of the seed cone of Douglas-Fir following dormancy. Can. J. Bot. 43: 317-332..

Pilger, R. 1926. Gymnospermae. In Engler, H. G. A. and K. A. E. Prantl (ed.). Die naturlichen Pflanzen- familien. 13. Leipzig. ' * / Radais, M. 1894. Contribution a 1'etude de l'anatomie comparee du fruit des coniferes. Ann. Sci. Nat. Bot.Paris 19: 165-369.

Riggs, E. S. 1926. Fossil hunting in Patagonia. Nat. Hist. N. Y. 26: 536-544..

Robison, C. R. and C. N. Miller, Jr. 1975. Glycol methacrylate as an embedding medium for lignitic plant fossils. J. Paleont. 49: 559-561.

Schopf, J. M. 1943. The embryology of Larix. Illinois Biol. Monogr. 19: 1-97.

Seward, A. C., and S. O. Ford. 1906. The Araucarieae, recent,-and extinct, Phil. Trans., London (B) 198: 305-411.

Shaw, F. J. F. 1909. The seedling structure of Araucaria bidwillii. Ann. Bot. 23: 321-334.

Silvicultural Notes on Forest Trees of Queensland. 1917. Qd. Forest. Bull. Dept, of Public Lands. No. 3. Part 1. 8-11.

Spegazzini, C. 1924. Coniferales fosiles Patagonicas. . Ann. Soc. cient. argent. Buenos Aires. .98: 125-139.. 140

Stoekey, R. A. 1975a.. Seeds and embryos of Araucaria mirabilis. Amer. J. Bot. 62: 856-868.

______. 1975b. Immature ovulate cones of Araucaria mirabilis. • Bot. Soc. Amer. Abstracts” Allen Press, Lawrence, Kansas, p. 25.

Takhtajan, A. L. 1953. Phylogenetic principles of the system of higher plants. Bot. Rev. 19: 1-45.

Thomson, R. B.. 1905a. Preliminary nojte on the Araucarin- eae. Science. (M.S.). 22: 88.

______. 1905b. The megaspore-membrane of the gymnosperms. Univ. Toronto Studies. Biol. Ser. 4: 85-146.

______\ 1907. The Araucarieae - A 'Proto-Siphonogamic* method of fertilization. Science. (N.S.) 25: 271- 272.

______. 1913. On the comparative anatomy and affinities of the Araucarineae. Phil. Trans. Roy., Soc. Lond. B: . 204: 1-50.

Tubeuf, C. von. 1892. Beitrag zur Kenntnis der Morph- ologie, Anatomie und Entwicklung des Samenflugels bei den Abietineen. Ber. bot. Ver. Landshut. 12: 155-196. * Wieland,.G. R. 1929. The world's two greatest petrified forests. Science. 69: 60-63.

Wilde, M. H., and A. J. Eames. 1948. The ovule and "seed" of Araucaria bidWilli with discussion of the of the genus. I. Morphology. Ann. Bot. (N.S.) 12: 311-326.

, and ■ . 1952. The ovule and "seed" of Arau- caria~bidwilli with discussion of the taxonomy of the genus. II. Taxonomy. Ann. Bot. (N.S.) 16: 27-47.

______, and . 1955. The ovule and "seed" of Arau­ caria bidwillii with discussion of the taxonomy of the genus. III. Anatomy of multi-ovulate cone scales. Ann. Bot. (N.S.) 19: 344-349.. 141

Windhausen, A. 1924. Llneas generales de la constitucion geologica de la region situada al oeste del Golfo de San Jorge. Bol. Acad. Cienc. c6rdoba. 27: 167- 320.

______. 1931. Geologia Argentina. 2'. Buenos Aires.

Worsdell, W. S. 1899. Observations on the vascular system of the female "flowers" of Coniferae. Ann Bot. 13: 527-548.