107
275

 

VASCULAR ANATOMY OF THE
SAPODILLA FLOWER
(MANILKARA ZAPOTILLA (JACQJ GILLY)

Thesis for the Degree of M. S.
MICHIGAN STATE COLLEGE

Marie Leonora Ehrmann

1950

 

 

This is to certify that the

thesis entitled

"VASCULAR ANATOMY OF
THE SAPODILLA FLOWER
(Manilkara zaootilla (Jacq.) Gilly)"

presented by

Marie Leonore Ehrmann

has been accepted towards fulfillment
of the requirements for

 

_M;_S_.__degree in Botany

3 E ‘5 20—1722: “1 ‘2

Major professor

Date @0045; 5" 151.51

0-169

VASCULAR ANATOMY OF THE SAPODILLA FLOWER
(Manilkara zapotilla (Jacq.) Gilly)

by
MARIE LEONORE ERMANN

A THESIS
Submitted to the School of Graduate Studies of Michigan
State College of Agriculture and Applied.Science
in partial fulfillment of the requirements
for the degree of

MASTER.OF SCIENCE

Department of Botany and Plant PatholOgy

1950

 

Efiiilg

TABLE OF CONTENTS

1. Acknowledgments...............................
2. Introduction.................................. 1
3. Review of Literature
a. Relation of floral anatomy to morpholosy
and taxonomy............................. 2
b. Systematic position of the Sapodilla..... 4
c. Histological and anatomical techniques... 8
4. Investigation
a. Plant materials utilized................. 11
b. Methods used in anatomical studies....... 11
c. Discussion of methods.................... 16
5. ObserVations.................................. 20
6. Interpretation and.Discussion................. 29
7. Summary.:..................................... 35
8. Conclusions................................... 37

90 Plates with Figureae000.000.eeeeeoeeoeeeeeoeee 39
100 BibliographyOOeeee0.00000000000000000000000000 43

‘7‘: i 3 2‘3. ‘3 73. 1

AQQTOWLEDGMENTS

For assistance, as specified below, the author would
like to express her appreciation to the following members
of the staff of the Department of Botany and Plant Pathology
of Michigan.State College: Dr. E. Fo‘Woodcock, for his
supervision and constant aid in histolOgical and anatomical
technique; Mr. G. L. Gilly, for his supervision of the
systematic (taxonomic) phase of this work, for his contin-
uous encouragement, and for supplying the materials studied;
Dr. G. B. Wilson, for advice on histolthical technique and
for supplying reagents used in some of the experimental
hist010gica1 work involved in this study; and Dre. w. B.
Drew and C. A..Alex0poulos for general helpful suggestions

during the course of this proJect.

INTRODUCTION

The family Sapotaceae includes a group of plants widely
distributed in the trepics and subtrOpios. This family is
characterized by a considerable variation and complexity of
floral structure; this diversity, resulting from the existence
of several evolutionary trends of reduction and fusion of
various parts and of various whorls in the flower, has led
to conflicting interpretatiom of the floral structure of
some of the members of this family. To date, these inter-
pretations have been based solely upon observations of the
gross morphology of the flowers. So far as can be deter-
mined, no previous study of vascular anatomy has been made

in flowers of the Sapotaceae.

The principal purpose of this study is an attempt to
prove, through investigation of the vascular system, the
proper designation of each whorl of parts in the flower of
the Sapodilla, Hanilkara zapotilla (Jacq.) Gilly, one of the
economically important members of the family Sapotaceae. A
secondary objective of the study is to provide additional
data of use in determining the systematic position of the
Sapodilla complex of species within the family.

Taxonomic and anatomical workers sometimes differ in their
opinions regarding the importance or the validity of basing an
interpretation of floral structure upon the vascular system of
the flower. However, it is believed that a study of the vas-

cularization is, at the very least, a valuable supplementary

tool in cases where morphological features alone fail to
give a complete or correct picture of floral structure.
'With this idea in mind, it is haped that this investiga-
tion will throw light on the flower structure of the
Sapodilla and related plants and.thus prove of assistance

in the interpretation and classification of the Sapotaceae-

REVIEW OF LITERATURE

gelation of Floral.Anatony to Morphology and.gaxonomy

 

Work has been done by, numerous authors to demonstrate
relationships between plant families, and subordinate groups
within families, on the basis of anatomical studies of the
vascular system in the flower. It has been shown that the
old classification of Dicotyledons along two specific lines,
Polypetalae and GamOpetalae, is invalid, since some members
of the two groups can be shown to be closely related on
basis of floral anatomy, though they are morphologically dif-
ferent with respect to the extent of petal fusion (Douglas,
1936). The families Primulaceae and Garyophyllaceae may serve
as examples of families in which both apopetalous and sympet-

alous members occur.

The phy10genetic origin of a floral organ may be indi-
cated by tracing its vascular tissue to its level or point of
origin he the main vascular cylinder of the floral axis. For
instance, according to Tillson (l9#0), it is indicated that

the scales found between the stamens and carpels in the flower

of Grassula sp. are derived.from the gynoecium, since they
receive branch traces from the dorsal bundle traces and often

from the median lateral bundle traces of the carpels.

There is, however, disagreement as to the validity of
using the vascular system as a guide to phlegenetic history.
Grelot (quoted by Arber, 1933) states: ”When there is partial
abortion of an organ, one finds ordinarily one or more rudi-
mentary bundles, which lead to it; when there is total abor-

tion there are never any bundles.“

Hunt (1937). on the other hand, makes this statement:
"It has been found that ancestral conditions of a structure,
long since lost to external view, may often be represented
by a lingering vascular supply. Plant vascular tissue, like

the skeleton of animals, is slow to change.“

Van Tieghem and Henslow (quoted by.Arber, 1933) described
bundles branching off the calyx midribs in.some members of the
Primulaceae, as representing an absent series of stamens.
Arbor, in disagreement, states: "Evidence is absent for the
survival of vascular tissue after the organ which it supplied
ceased to exist. [Parthenon] an organ, which retains some
trace of its normal external form, may yet show complete lack
of vascular tissue.”* On the other hand, she states that the

androecium is particularly favorable to anato-morpholOgical

 

* The author of the present paper found this to be the case in
the glands at the base of the gynoecium.in Grassula cogymbulos .

4

study, since abortive stamens occur more frequently than

rudhments of other floral parts.

In general, sepals are essentially leaf-like in their
venation, whereas the vascular pattern of petals resembles
that of sterile stamens; i.e. there are usually one to three
(sometimes more) prominent veins in a petal. According to
Eames (1931), however, petals in their vascular supply may
often be like leaves. He states ”Petals appear to be some-

times modified leaves, like the sepals.....".

Swamy (1949) reports a striking similarity in the method
of vascular supply of sepals and.outer petals in members of
the family Degeneriaceae. Each petal and each sepal of
Degeneria sp. is supplied by one median, and two marginal
traces. The inner petals of the flower receive one trace,

which forks into a median, and two marginal traces.
Systematic Position of the Sapodilla

The Sapodilla (Manilkara zapotilla (Jacq.) Gilly; the
Achras zapota of most authors) is a member of the family
Sapotaceae; this family is generally included in the order
Ebenales. The Sapotaceae is a family of tropical and sub-
tropical woody plants (mostly trees); some members of this
family produce latex. Some types of this latex are used
commercially in the production of various gums, such as
balata rubber, chicle, and.gutta percha. The leaves are

usually simple, entire, and of coriaceous texture. .Arrange-

ment is alternate or rarely Opposite. The perfect (or rarely

5
polygamous) flowers are solitary in the leaf axils, or borne
in axillary clusters. They are generally described as being
sympetalous, with epipetalous stamens in one to three whorls.

The authors are extrorse, and petaloid staminodes are fre-

quently present in the flowers.

The ovary may consist of as many as fourteen or as few
as one, uni-ovulate carpels. Placentation is essentially
axile. The style is sometimes obscurely lobed at the summit.
The fruit is a large, generally edible, berry; the seeds are

variable in.shape and.structure.

The genera of the family Sapotaceae have been variously
grouped by taxonomists (Baillon, 1892; Dubard, 1915; Lam,
1925, 1927. 1939, 1941). One of the more recent classifica-
tions is that of Lam (1939), in which the family is subdivided
as follows: ‘

Subfamily I. Sideroxyloideae, divided into three tribes:
Sideroxyleae, Bumelieae, and Pouterieae;

Subfamily II. Mimusopoideae, divided into two tribes:
Mimusopeae and Manilkareae;

Subfamily III. Hadhucoideae, divided into two tribes:

Madhuceae and Palaquieae.

This classification is based primarily upon presence or
absence of appendages on the corolla, the number of members
in the perianth, the general type of seed scar, and the
presence or absence of staminodes. In this classification,

the Sapodilla is placed in a group FAcyclicae" of the subtribe

Pouteriineae under Tribe Pouterieae of the Sideroxyloideae.
The rather obviously related balata producing members (M-
ggg spp.) of the family are placed in the subfamily Mimuso-
poideae, while the gutta-percha producing plants are placed
in subtribe Palaquieae of the Madhucoideae.

More recently, Lam (1941) has included the Sapodilla
in the subtribe Manilkareae of the Mimusopoideae, under the

generic name Achras L.

Cronquist (1946), in a cons ideration of the North American
members of the Sapotaceae, seems to be in agreement with Lam
as to characters used in recOgnizing subdivisions of the family.
The presence or absence of staminodes, presence and position
of appendages on the corolla lobes, type of seed scar and of
ovule attachment, and the presence or absence of endosperm, are
generally important features. He states, however, that number
and arrangement of parts per whorl often fail in this respect.
Like Lam (1941), Cronquist places the Sapodilla in close re-
lationship with the balata producing members of the MimusOpoi-
deae. Unlike Lam, however, and following the treatment of
Gilly (1942, 1916), Cronquist places the Sapodilla in the genus
Manilkara Adana .

The subfamily Himusopoideae is generally characterized by
cyclic flowers and by possession of paired dorsal appendages
subtending the corolla lobes. The genera in this subfamily are
grouped (last, 1939, 19“) into two tribes, the MimusOpeae and

the tianilkareae; the former are characterized by having small,

7

basilateral seed scars, and eight members (in two groups of
four) in their floral whorls. Mimusgps L. (which is fre-
quently divided into several sections), Baillonella Pierre,
and.Butygospermum.Kotschy are the more important genera or
this tribe.

The Manilkareae is differentiated from the MimusOpeae
by the elongated, generally lateral seed scars, and six
members (in two groups of three) in the floral whorls. The
principal genera in the Manilkareae are Manilkara.Adans.
(which includes the sapodilla and the balata), Muriea HartOg,
Northia Hook.f., Shaferodendron Gilly, and several other small
genera of the eastern hemisphere. The most striking feature
of this tribe is a tendency towards reduction of dorsal sp-
pendages and staminodes. .According to Lam (1941), these
reductions occur independently of each other. In addition,
Gilly (1942) has shown that the paired dorsal appendages may
be partially or entirely fused with the petals.

The Bapodilla was originally placed in.Achras by Linnaeus
(1753). This genus was separated by subsequent workers
(Eichler, quoted by Lam, 1938; Engler, 1903; Dubard, 1915;
Lecomte, 1920; Lam, 1925, 1927, 1938, 1939, 1941) from Egali-
3233 only because, in the former, the dorsal appendages were
completely obsolete, perhaps through fusion with the petals.
In manilkara the appendages were always present. Gilly (1942)
has shown, however, that a large number of intermediate forms,

exhibiting various degrees of union between the "dorsal (or

stipular) appendages" and the ”petals", exist; these range
from conplete separation (as in the balata flower) to com-
plete fusion (as in the Sapodilla flower). For this reason
he incorporated the Sapodilla complex into Menilkara; the
nomenclaturally technical reasons for the choice of generic

name were discussed in detail by Gilly (1943).
Histological and.Anatomical Technigues

The two most common methods used in the initial process
of preparing dried flowers from herbarium specimens are (a)
soaking in water and (b) boiling in water. One author working
with buds from herbarium sheets (Tillson, 1940) reported
soaking specimens in water at 80 degrees a. for two hours
prior to bleaching in 5% NHAOH to be satisfactory for soften-
ing the buds of Kalanchoe and.Bgzgphyllum. Boiling a few
minutes in tap or distilled water has been recommended by

,others.

To dehydrate specimens in preparation for clearing and
embedding, a series of ethyl alcohol solutions of increasing
concentrations is generally used. This process, in addition,
necessitates gradual transfer of the material from absolute
alcohol into pure xylol, since paraffin, the commonly used
embedding matrix, is soluble only in the latter.

The use of xylol as intermediate agent between dehydration
in ethyl alcohol and.paraffin embedding may be eliminated'by
applying Zirkle's Butyl Alcohol.Method (Rawlins, 1933). since

paraffin is soluble in this reagent. Since butyl alcohol is

slower in penetrating the tissues, it is preceded by ethyl
alcohol, then gradually mixed in. (This method is described

later in this paper.)

A modification of this method consists in setting the
material into a 10% glycerin solution, allowing the water to
evaporate, and than transferring to a 50% glycerin-50% normal
butyl alcohol solution for 36 hours. This is followed by
three changes of pure normal butyl alcohol, 36 hours each,

then by infiltration with paraffin (cited in Rawlins, 1933) .

Ethyl alcohol may be substituted for butyl. Three changes
of 95% ethyl alcohol, three hours each, are followed by two
changes, one to two hours each, of absolute alcohol. This is
much less time consuming than the use of butyl alcohol, but
necessitates transfer into xylol, benzol, or a similar clearing

agent.

Numerous materials are used for the purpose of clearing
plant tissues after dehydration. Generally they belong in
either of two groups: the hydrocarbons, such as xylol, benzol,
toluol, phenol, chloroform, etc., and.the essential oils,
including clove oil, cedar oil, bergamot oil, aniline oil,
oil of wintergreen, and others. Vegetable oils cause the least
amount of shrinkage (Lee, 1928).

The above described clearing agents, however,*will not
remove brown pigments of dried plants, and thus, for treating

herbarium specimens, definite bleaching agents are required.

10

NHADH,‘NaOH, lactic acid, peroxide, dioxan, chlorine, and
sulfurous acid, and.others, are listed as bleaching agents
by various authors. (Methods utilizing the first four
bleaching agents mentioned are described later'in this
paper.) It must be remembered that plant tissues vary in
chemical composition and physical properties, and that this,
as well as the treatment the tissue has previously undergone,
will necessitate the use of different bleaching agents for
different plant materials.

Several stains were recommended by various authors to
differentiate vascular tissue. According to Johansen (19#0),
a saturated solution of light green, aqueous or alcoholic,
acidulated with hydrochloric or acetic acid, will stain lig-
nified tissue, and wash out from the remaining tissue. The
same author recommends a one percent solution, in 70% alcohol,
of Iodine Green. Other stains in use for this purpose are
Methyl Green, Safranin.0, Crystal Violet, and.Eosin. McClung
(1937) suggests a one percent solution of Safranin 0 in 50%
alcohol, the alcohol solution being prepared with "aniline
water“. The latter reagent consists of two parts aniline oil
per 100 parts water. Combinations of a) Methyl Green and
Bismarck‘Brown, b) Gyanin and.Erythrosin, and c) Safranin

and Light Green are suggested by'Woodcock (1945).

11
INVESTIGATION

Plant Mategials Utilized

Flowers from the following herbarium specimens were

used.in this study of the vascular anatomy of the Sapodilla,

and related plants:

 

Species Specimen Lgcality
Manilkara za otilla
(Jacq.) Gilly Alexander s.n. Jamaica
' Box 1210 Antigua
Broadway s .n . Grenada
Duque 17#7 Colombia
Eyerdam 212 Haiti
Harris 8634 Jamaica
Jack 7820 Cuba
Lundell 500 British Honduras
Matuda 3254 Mexico
Pittier 779 Colombia

Ricksecker 316
Rose and.Rose 21609

Hanilkara bidentata
A0 DC. Chev. DUES 3263
Manilkara bahamensis
Baker) Lam & Meeuse Curtiss 5437
' Hitchcock s.n.
Seibert 1317

Manilkara tapggaensis

illy 'Woodson,.Allen, and

Seibert 1455

Virgin Islands
Curacao

Guadalupe

United.States
Bahama Islands
Uhited.8tates

Panama

 

ggthods Used in Anatomical Studies

To facilitate explanation and discussion of techniques

used in this study, all methods are outlined in the following

pages, and.numbered or lettered.

The time factor was important

12

only in certain steps of each method, and therefore, is

not given in all instances.
De dration

I. Ethyl Alcohol Method

35% 31°Oh01eoeeeeeeeeee002 " A> hours Minimum

20 50% aIOOhOIOOOOeeeeeeeeeo “
30 70% aICObOIOeeeeeeeeeeeee 9 n :
A’e 85% 81c0h°1eeeeeeeeeeeeee :0 2 2

5e 95% alcohol....o.........
6.100% alcohol.............. 2 hours minimum
7. Change of 100% alcohol.... ”

80 Change Of 100% a1°°h°1eeee fl :. ...:

II. Zirkle' e Butyl Alcohol Method, using Normal Butyl Alcohol

1. 95% water, 5% ethyl a1c0hol --- ... one hour
20 89 u 11 3. " --- ee e I u
I I I. I I
it 33?; .. ' £2 .. .. z: 2:: .~. .~.
5. 50% " 40% " " 10% butyl alcohol“ "
6. 30% II mx '0 W 20% N 00
7. 15% " 50% " " 35% !' '9 1 hr.minimum
8. 5% N 40% M I 55% I I N
9- --- ._ 25% 9 9 75% 9 ' 3 a 3
10. --- --- - 100% v v V v "
11. Change of 100% butyl alcohol 3 : 2

12. Change of 100% butyl alcohol

III. Same technique as II, but combining steps 2 and 3, and
substituting 85% water and 15% ethyl alcohol.

IV. Same sequence as II, using Tertiary Butyl Alcohol instead
of Normal Butyl Alcohol.

V. Glycerin Method

1. Immerse in 10% glycerin solution, allowing water to
evaporate .......(approx. 31 hours at 60° C)

20 95% ethyl 81301101 Oeeeeeeoeee0003 hour.

3. Change of 95% ethyl alcohol....3 hours

4. Change of 95% ethyl alcohol....3 hours

l3

Clearigg and.§1eachigg

Ae xy101 “Othod

lo 100% ethyl alcohol.o...........oo2 to 6 hours
2. 3 parts ethyl alcohol, 1 part xylol

3. 2 ' ' ” 2 parts ”
4. 1 part 9 9 3 '
5. 100% xylol. _

6. Change of 100% xylol

7. Change of 100% xylol

B . Benzol Method
Same sequence as in.method A, substituting benzol for xylol.
C. .Ammonium.Hydroxide Method

1e ‘Watar at 60° CeeeeeeeeeeeeeeeeeeeeeeG hours
2. 5% solution of NHqOH at 60° C.......overnight
3. Rinse in.running water bath.........l2 hours minimum

D. Same sequence as 0, using 10% strength of NHhoH
E. Sodium Hydroxide Method

Same sequence as method 0, substituting 5% NaOH for’NHAOH
F. Same sequence as method C, using 10% NaOH instead.of‘NH40H
G. Peroxide Method

1. 95% ethyl alcohol
2. 95% ethyl alcohol and 3% peroxide at 60° c...24 hrs.
' minimum
3. As many changes of above step as required
for complete bleaching of tissue
(variable)0..00O.OOOOOOOOOOOOOOOOOOOO.000.024 h".

each
4. 95% ethyl alcohol
5. One or two changes of 95% ethyl a1c0h01

H. Lactic Acid Method

1. ‘Water (distilled)
2. 75% lactic acid......................several days
3. Wash in running water bath...........2# hours

J. Clove 011 Method

1. 100% ethyl alcohol
20 01°" 01]., at 60° CeeeOOOeeoeoeoeeoooa days M1111.“
3- 100% ethyl alcdhol.

14

K . Cedar 011 Method

1.
2.
3.
4.
5.

100% ethyl alcohol

2 parts ethyl alcohol, 1 part cedar oil...26 hours
1 part 2 part! I N ee02 day.
100% ethyl alcohol........................24 hours
One change of 100% ethyl alcohol..........”

L. Chloroform Method

1.
2.

3.

100% ethyl alcohol
100% ethyl alcohol and chloroform, equal

parts, at 60° 0......o..................2 days
100% ethyl alcohol

Staining of Sectioned Material

1) Methyl Green- Bismarck Brown

1.
2.
3.
4.
5.
6.

Methyl Green (aqueous)............overnight

waterOOOOOOOOOOOOOOOO0.000.000.0002 minutes

Bismarck Brown, 2% (70% alooholic)l/2 minute
95% ethyl alc°h°1eeeeeeeeeeeeeeeeea

100% .................2 minutes

Xylol

Staining of Bulk Material

1) 25% Alcoholic Eosin

1.
2.

3.
4e

StaineeeeeeeeeeeeeeeeeeeeeeeeeeeeeapprOXe 2 dflya
Destain in 35% alcohol (ethyl) or

35% acid.alcoh01
3g; .thyl aIOOhol.................f:' minthB

2) 1% Aqueous Eosin

1.
2.
3.
4.
5.

Stains.eeeeeeeeeeeeeeeeeeeeeeeeeeeapprOXe 10 hour.
'3tor.............................fe' unUtQB
70% athyl aloOhol.................

95 ...OOOOOOOOOOOOOO ' u

100% " u

38)

3b)

46)

4b)

5)

15

1% Aqueous Saffranin 0

1.
2.

(Ii-4 mung-u

StaineeeeeeeeeeeeeeeeeeeeeeOeeOaPPPOXO 10 hours
watorCC0.0......OOOOOOOOOOOOOOOfg' minutes

70% ethyl alcohol..............

70% acid ethyl alcohol.........ug to 5nminutes
Change of 801d alcohol.........

70% ethyl alcohol..............15 minutes

95% ethyl a1cohol..............

100%

50%.Alcoholic Saffranin.0, the alcohol solution

consisting of 'anailine water“ (approx. 1/3
aniline oil and.2/3 water) and 95% ethyl
alcohol, equal parts...........approx. 2 days

m% acid ethyl alcohol......... " minutes
70% .........10 to 15 minutes
95% ethyl alcohol..............few minutes
100% “

Aqueous Iodine Green

StameeeeeeeeeeeeeeeeeeeeeeeeeeapprOXe 2‘ hours
Water.......................... 2h°ur.
50% ethyl alcohol..............rev minutes

70% acid ethyl alcohol.........few hours

70% ethyl alcghcl..............10 minutes minimum
300% .. .. IIIIIIIIIIIIII! ~ ~
100% " ” and xylol, equal parts

70%.Alcoholio Iodine Green

1.
2.
3.
#.
5.

Stain..........................appr0x. 2 day!
70% Cthyl alCOhOleooneoeeoeeOOlo m1HUtea

70% acid ethyl alcohol.........10 to 1? minutes
95% ethyl alcohol..............! ”
100% ' .

1% Aqueous Cryltal Violet

l.
2.
3.
4.
5.
7.
8.
9.

10.

Stain..........................appr0x. 10 hours
Water..........................10 minutes minimum
50% ethyl alcohol..............”

70 ..............15 minutes minimum
70% 301d ethyl 8.].0015101eeeoeeee0u

70% ethyl 310°h°1eeeeeeeeeeeeee ... I I
95% ..............10 minutes minimum
100% I O1 eeeeeeeeeeeeee3 mmUtea

100% ethyl alcohol and xylol,
equal parts...............few minutes
xylol '

, 16

6) 1% Aqueous Methyl Green

10 Stain..........................1 to 10 hours
2' Water-conuu.................10 minutes

3. 70% ethyl alcohol.............."

4. 70% acid.ethyl alcohol.........l. hour*minimum

5. 95% ethyl alcohol..............few minutes

601.00% ’ " 000.000.0000.. u n

'7. 100% ethyl alcohol and.xylol, _
aqua]. part8............... N

8. Xylol

Discussion of Methods

All herbarium flowers were softened by boiling very
gently for ten to twenty minutes in water. Both distilled
and tap water were used in order to determine the effect
of each upon solubility of the brown pimsnt of the dried
flowers during the boiling process. No difference in effect
was noticed, neither type of water removing more than a
negligible amount of the color. In some specimens single
parts of the perianth, and in one or two cases, the so-called

corolla tube became detached in the boiling process.

No difference in efficiency of dehydration was noted in
comparing the ethyl alcohol, butyl alcohol, and glycerin
methods. Experimentation with these methods was carried out
on freshly cut Crassula, Saintpaulia, Verbena, and Solanum
flowers. It was observed that the butyl alcohol methods
(numbers ll, III, andev) seemed to render these flowers more-
brittle than ethyl alcohol; thus these methods were not em-

‘ ployed in the work with sapodilla flowers. No differencein
effect was evident between Normal and Tertiary butyl alcohols.

Because of availability of materials, and closer familiarity

17

with the method, the ethyl alcohol method (number I) was
used, in preference to the glycerin (number V) method, in

dehydrating the Sapodilla flowers used for study.

By experimenting on Sapodilla flowers from herbarium
specimens, it was found that none of the ordinary clearing
agents (A, B, J, K) removed any amount of the brown pigment
(which was, at this point, almost entirely retained in the
specimens). Lactic acid (method.H), after a few days, re-
noved.some, but by far not enough, color to clear the flowers.
Chloroform (method.ld, NaOH, and.NH40H (methods 0 through F)
were superior to lactic acid, but still not strong enough.

The only reagent which completely bleached the flowers to a
transparent white color, and thus rendered them fit for stain-
ing, was peroxide (H202), applied as described in method G.
Rawlins (1933) suggests only one application of peroxide in
95% alcohol, but it was found that, to clear adequately the
dried Sapodilla flowers, at least three changes of H202 were
necessary. Two undesirable effects of this method were (a)
undue softening of the flower to the point where it practically
disintegrated upon attempt to dissect it or cut it into longi-
tudinal halves, and (b) release of oxygen, from the peroxide
solution within the tissues, which caused the flower to float.
The latter condition necessitated aspiration of the specimens
before paraffin infiltration. Two or three changes of 95%
ethyl alcohol subsequent to bleaching served to remove the
peroxide and.harden the tissue to a workable texture. The

specimens to be sectioned were then transferred through 100%

 

18

ethyl alcohol, xylol (method.A), and gradually infiltrated
with paraffin.

Apparently there is some air trapped.within the flower,
which cannot be removed'by aspiration, because the paraffin
blocks with the embedded Sapodilla material always contained
numerous air bubbles, especially in and.around.the gynoecium;
this often resulted in greater or lesser failure of ribbon
formation. (This difficulty was never encountered in working

with fresh Qrassula, Saintpaulia, Verbena, and Solanum flowers.)

The sections were attached to slides by egg albumen fix-
ative, the latter being made up in preportion of 50 cc. egg
albumen, 50 cc. glycerine, and one gram sodium salicylate
(Woodcock, 1945), then floated in two percent formalin. Upon
drying, the sections were stained by the method listed. They

were then mounted in.Canada Balsam.

Specimens which were studied.in bulk were run‘badh from
95% alcohol to the medium in which a particular’stain.was
soluble. At first, pieces of tissue, such as a sepal or petal,
and later, longitudinal quarters of flowers, were used to test

the six stains listed.

The Basin stains, both aqueous and 25% alcoholic, were the
least usable, and thus were discarded after a few trials.
There was an unequal distribution of the stain throughout the
tissue, with scarcely any differentiation of the vascular
bundles. Destaining in acid alcohol resulted.in.precipitation
of the stain.

19

The Saffranin solutions showed more promise than the
eosin, but were surpassed.by Iodine Green, which displayed
a better differentiation of the veins. In these preliminary
staining trials, stains #a and b were best; stains 5 and 6
were next best. The green and purple stains seemed in all

cases superior to the red ones.

The four'stains mentioned in the preceding paragraph
were repeated, each on one one-fourth of a sapodilla flower
(bleached in peroxide). In the final analysis, stains 5 and
6 were found most efficient; timing experiments showed the
Optimul.amount of time for each step to be as described.in
the preceding section of the paper. Methyl Green was slightly
superior to Crystal Violet, and.was finally used in the bulk
staining of all other flowers studied. One difficulty was
encountered.in all types of stains (though least in numbers
5 and 6); a much heavier stain would be retained in the lower
portion of the gynoecium, and in the receptacle, than in the
other floral parts, so that it was impossible in case of the
red.stains, and difficult with the most workable stains, to
strike a favorable balance between understaining the latter,

and overstaining the former.

The specimens stained and destained.in bulk were carefully '
laid open by making a shallow out along the length of the
peduncle and.the receptacle. Through careful manipulation it
was possible in.most cases to determine the relative levels

at which the various traces diverged from the main cylinder of

20

the peduncle into the floral parts, and to follow the entire
'course of each branch. Two of the flowers were badly dis-
integrated by the peroxide, and two or three others did not
take the stain well enough for conclusive study, but the
partial results that could be obtained from them appeared
to be in accordance with the vascularization picture pre-

sented by the others .

Longitudinal sections, 14 microns in thickness, were
cut from one flower (Alexander s.n.) and transverse sections
of the same thickness were out from another flower-(Broadway
s.n.) . Since the flowers were at a stage of maturity where
the sub-epidermal cell layers of the ovary had already become
lignified, extreme difficulty was encountered in cutting
paraffin ribbons . Sufficient sections were obtained from
each flower, however, to include the area where differential

vascularization into the floral whorls occurred.
OEERVATIONS

Because of disagreement in the interpretation of the
floral structure of the subfamily MimusOpoideae, the termin-
ology of the different floral whorls varies with different
authors. To avoid possible confusion in reporting observa-
tions, each whorl will be designated by a letter, correspond-
ing to the floral diagram, Figures 1 and 2. The following
tabular arrangement shows a comparison of the nomenclature
used for each whorl by previous workers (see Review of Liter-

ature) and by Gilly (191}2) .

21

Whorl Previous Workers ' c1111

A Sepals, outer series Outer perianth,
valvate (sepals)

B Sepals , inner series perianth Inner perianth,

imbricate (petals)

0 Dorsal (stipular) appen perianth
dages to petals (in xterior staminodes
Balata complex) (0 and D are fused
in the Sapodilla,
D Petals separated in the
' Balata flowers)
E Staminodes androecium Interior staminodes
androecium
F S tamens S tamgng
G Carpels (ovary). . . . . . . . . .gynoecium . .Garpels (ovary)

 

The vascular skeleton of the flowers that were laid open
as shown diagrammatically in Figure 3. The first three traces
which separate from the vascular cylinder, at the lowest level,
form the midribs of the three segments of whorl A; two: of these
three tracesere indicated in black in Figure 3. These traces
closely parallel the cylinder for some distance and, after
diverging into the base of the segments, each gives rise to
two branches which parallel the main trace toward the apex of

the segment .

At the next level of divergence, three strands separate.
Two of these (stippled in Figure 3) make their way also into
the sepal, one above the other, and there give rise to several
veins, each of which branches again to form the conducting
system of this segment .*

 

*Another case, where the lateral sepal traces arise
independently of the midrib is described by Tillson (1940) in
some members of the Kalanchoideae (subfamily of the family
Ericaceae); there the laterals separate at a lower level than
the midrib. The family is close in taxonomic position to the
Sapotaceae. ,

22

The third strand (shown by the horizontal lining in Figure 3).
passes into a segment of whorl B as a midrib. At the base of
this strand.two branches arise; each dividing several times to

form the vascular'supply of this segment.

The veins entering the combined whorls 0-D branch off
independmtly and at a much higher level, as shown by the
diagonally striped lines in Figure 3. In making this diagram,
a petal was removed to expose the vascular system, and only
a part of the structure formed by whorls C, D, E, and F, is

shown.

The venation pattern of the lobe (whorl C-D) to which the
stamen is attached, and that of the staminode (whorl E) are
essentially the same, both structures generally containing
one strong vein that branches into two parts. The midribs of
each member of whorl 0-D and the adjacent member of whorl E.
originate from the same strand at the base of the tube. The
single trace of the filament (whorl F) diverges as a third
branch from the midrib of each segment of whorl 0-D. The
exact point of divergence varies somewhat from.flower to
flower and to a greater degree from species to species, but

is generally close to the base of the tube.

The last two groups of traces to separate from the central
cylinder are those of the gynoecium (whorl G). The dorsal
carpel traces separate first and continue through each carpel
and upwards within the style. The two lateral traces of each

carpel are apparently fused into a single ventral bundle and

23

these ventral bundles terminate below the middle of the ovary.
The style is hollow afia internally lobed (the number of lobes
corresponding to the number of carpels) throughout its length.

The parallel broken lines in Figure 3 correspond to the
levels at which cross-sectional diagrams (Figures 4 to 10)
were drawn. The lowest cross section.(Figure 4) shows three
strands diverging from the main central cylinder; these
' evidently represent the midribs of whorl A. The separation
of these traces apparently takes place at a relatively low
level in the peduncle; at least it was below the point at

which the first sections of this flower were cut.

The next diagram (Figure 5), approximately 550 microns
above the first, shows that three more bundles, alternating
with the first three, have separated from the main cylinder.
In'rigure 6, each of these additional three bundles has
divided once to form a pair. One member of each pair divides
again, so that three groups of three bundles are formed.
This is shown in Figure 7. In each of these groups, two
traces apparently represent the lateral traces of whorl A
(represented by stippling in Figure 3), the third being the
midrib of whorl B (represented by horizontal lining in Figure
3).

The fourth cross section (Figure 8) represents a level
550 microns above the previous one, and shows the divergence
of the six traces which enter the tube formed by the fusion
of whorls C, D, E, and F.

In the next diagram (Figure 9) a cross section through

24

the lower part of the ovary is shown; in this flower the
number of carpels is eleven. One dorsal and one ventral
trace (formed of the two united ventral bundles) per carpel
are evident. The last cross section (Figure 10) includes a
section through the style, showing its internal structure,
and the dorsal bundles, which are its sole vascular elements.
The large number of traces seen in each of the surrounding
floral whorls indicates the leaf-like character of venation
displayed by all floral organs. Only one trace is present
in the anther. The extrorse anthers are turned at an angle
from their natural position, evidently to fit into the lobed
shape of the tube resulting from compression of the flower

in the embedding process.

The diagram, Figure 11, representing a longitudinal
slice through a saDOdilla flower near the center, was drawn
from ten successive sections, each 14 microns thick. Since
the section is not a perfect median one, a small portion of
a segment of whorl A is present on the right side. The
portion of a vascular trace in the lower part of this tissue
is probably a part of the midrib; the trace passing into the
upper tissue of the segment is evidently one of the laterals,
since it diverges from the central cylinder at the same point
as the trace extending into the segment of whorl B.

The trace entering the left-hand segment of whorl B is
probably the midrib of that segment, Judging from its level

of divergence.

25

At a much higher level (the same level on both sides),
a single trace departs from the main cylinder to enter the
staminode (whorl E) on the left side, while a similar trace
on the right passes into a segment of whorl 0-D as a midrib.
This trace divides, the left branch extending up through the
stamen (whorl'F), the right arm continuing upward within
segment 0-D. Two dorsal carpel traces are shown, extending

upward within the hollow style.

The venation of individual members of the floraltwhorls
is shown in Figures 12-14. Figure 12 represents one of the
three valvate segments of whorl A; the three segments are
Joined at the base into a tube, and each segment is very thick
in the basal part. The midrib is embedded in the lower part
of the tissue, while the two lateral traces are situated near
the upper epidermis. As may be noted from the drawing, the
two lateral bundles are thicker than the midrib, and, by
branching, each gives rise to the vascular supply of one half

of the segment.

One of the three segments of whorl B is represented by
Figure 13; these three segments are imbricate and independently
attached upon the tube formed by whorl A. They are somewhat
more pointed than the segments of whorl A, and not so thickened
at the base. The vascular supply of each of these segments
originates from the single prominent trace which gives rise to
the midrib and.a branch on each side; several veins diverge

from these branches and run almost parallel to the midrib.

26

The venation pattern of the segments in both whorls A
and.B is essentially leaf-like, and may be described as
flabellate. It is very similar in both series, the only
difference being the presence of the two lateral veins in

the outer one.

The basic vascularization of segments of the united
whorls C, D, E, and.F is shown in Figure 14. There is always
one prominent midrib in each segment of whorl 0-D, and gener-
ally one strong trace passing into each segment of whorl E.
The midribs of segment C-D and of one of the adjacent members
of whorl E are united at, or even below, the base of the tube.
As previously mentioned, one strand separates from the midrib
of each segment of whorl 0-D and extends into the stamen (a

segment of whorl F).

Though the basic pattern described above is constant in
the flowers of the Sapodilla and related plants, such as the
Balata, minor variations occur within a single species; some
of these are illustrated in Figures 15-21. Figures 15-18
show portions of the tubes taken from different specimens of
Manilkara zapotilla. Figures 19 and 20 represent tubes of
related.species (the former being from g. emarginata Gilly,

and the latter from g. EEQOgaensis Gilly).

In Figure 15, each branch of the midrib of each broad
segment (representing whorls C-D) is almost immediately divided
again to form a pair of branches; the inner one of each pair is
slender and extends toward the tip of the segment. There is,
therefore, no true midrib within the segment.

27

A somewhat different type of venation in whorl C-D is
shown in Figure 16. The midrib forks at or near the base of
the tube, and one of the branches divides almost immediately,
so that there are three principal veins extending into the
segment. The other branch divides too, but at a much higher
Ilevel. The venation of the staminode (whorl E) in this spec-

imen is quite weak.

An interesting example of accessory venation occurs in
the specimen shown in Figure 17. In addition to the midribs
of the broad segment and the staminode, there is (in four or
five pairs out of the six) on one or both sides of the staminode
midrib, a rather prominent trace, which forks near the summit
of the tube and sends one branch up into the stamhnode, the
other into the adjacent broad segment. These prominent acces-
sory veins are Joined to the staminode midrib at the base of
the tube. In one of the six broad segments, one of the branches
of the midrib diverges at a wide angle before forking; one arm
of the fork extends upward along the edge of this segment, the

other enters the adJacent staminode.

This extra venation seems to predominate in those forms
of Manilkara zapotilla, where the segments of whorl 0-D are
very Slightly three-lobed; that is, apparently the fusion be-
tween segments of whorl 0 and those of whorl D'is not so
entirely complete as in the Specimens where the broad seg-
ments (whorls 0 and.D combined) are entire. Only in one
specimen of the latter type was evidence found of prominent

venation.

28

In the specimen shown in Figure 18, no accessory prominent
veins are present, but there is a distinct network of small

veinlets connecting the broad segments and the staminodes.

In Manilkara emarginata, E. tabOgaensis, and fig. W,
in all of which whorls 0 and D are distinct (that is, each breed
segment has a pair of "dorsal appendages"), the venation pat-
tern is closely similar and.quite uniform. Segments of the
tube formed by whorls 0, D, E, and F of the first two species
mentioned are shown in Figures 19 and 20. In both species,
the venation of a staminode and.of a segment of whorl 0 are
very similar: the midrib of each gives off one branch on each
side, which diverges to pass into the adJacent "dorsal append-
age"; the central trace of each continues upward in the whorl
0 segment and forks once. Thus each "dorsal appendage" receives
traces from the members adJacent to it on each side, but the

greater part of its supply results from the branching of the

trace received from whorl C.

A very interesting feature observed in the Specimen of
m. emarginatg studied, was the fact, that there are three
distinct traces extending up through the filament. Time did
not permit sections to be made through the stamen, but from
superficial examination of the transparent anther under the
dissecting microscOpe, the venation of that organ appeared
to be as shown in Figure 19. The central one of the three
traces continues upward through the connective, but the other
two apparently each enter a pollen sac, and then immediately

become Obsolete.

29
INTERPRETATION AND DISCUSSION

This section of this paper is concerned with a con-
sideration of floral whorls A to F, inclusive, of the
sapodilla flower. Since the interpretation of whorl G
(the gynoecium) has never been in dispute, it will not be
considered here. 4An adequate description of the gynoecium

has been given in the preceding portion of this paper.
gerianthflwhorls A and B):

Though the venation pattern is similar in both whorls
A and B, the inner one is supplied by a different system
than the outer; in both there is one trace functioning as
midrib (the midrib of the inner series branching off at a
higher level than that of the outer), but each part of whorl
A is also supplied by two other traces separating from the
main cylinder at the same level as the midrib of whorl B.
In the segments of both whorls A and B, the midribs branch
near the base, giving rise to one vein to the right and one
to the left. ‘While in the segment of whorl B the minor
venation pattern originates from the branching of the midrib,
that of the segments of whorl.A originates primarily from
branches of the lateral traces entering the sepal at a level

above that of the midrib.

Furthermore, close observation of the so-called biseriate
calyx (whorls A and.B) reveals, that the outer three parts
(whorl A) are valvate in position, and have acute apices,

while the members of the inner series are imbricate, and have

30

more rounded, blunt apices. In addition, the three segments
of series A are united at the base into a short tube, and
the three members of series B are separately attached on the
tube formed by the union of the segments of A. The members
of whorl B are much less thickened at the base than are the

segments of whorl A.

Eames' statement to the effect, that petals may be
leaflike in venation, and.Bwamy's report of the similarity
in the venation patterns of the sepals and of the outer
petals of Qegeneria sp. (see Review of Literature) show,
that in some other plant families, the relationship between
sepals and petals is comparable to that existing between
whorls A and.B of the Sapodilla flower. The petals of the
flower of the Degeneria sp. receive one trace, which forks
into a median and two marginal traces. This situation is
the same as that found in members of whorl B in the Sapodilla

flower.

The above described results of previous morphological,
and present anatomical studies seem to indicate, that whorls
A and B (the so-called biseriate calyx) represent two distinct
perianth whorls: an outer series of three sepals (whorl A),
and.an inner series of three petals (whorl B). In the light
of this interpretation, members of whorl A shall, in the re-
mainder of this paper, be referred to as sepals, and segments

of whorl B, as petals.

 

 

 

31
Androecium.(!horls C, D, EI and.F):

The high level, at which the traces depart into the
base of the tube formed by the united whorls 0-D (the
structures considered originally to be corolla lobes)
indicates, that these structures are androecial in nature,
rather than corolloid. The venation of each member of
series C-D is similar to that of the staminodes (whorl E);
it is especially noticeable that, in the members of both
whorls, the midrib forks into two branches, each of which
again may branch. In general, the main venation pattern
of members of whorls C-D and.E tends to be dichotomous.
The single trace leading through the stamen arises as a
third branch which is separated, almost at the base of the
tube, from the midrib of the broad petaloid segment (whorl
C-D).

Since the basic pattern of venation is the same in all
Sapodilla specimens studied, the slight variations in the
levels, where the branches from the midribs (in both whorls
0-D and E) diverge and branch again, (as shown.in Figures
15-18) seem to be of minor importance. These variations are
not always constant in different segments of the tube of one

flower.

An interesting feature is the extra or accessory venation
found mostly in those forms of Manilkara zapotilla where the
somewhat three-lobed shape of each broad segment is apparently
suggestive of incomplete fusion of the dorsal appendages.

32
(whorl D in the Balata flower) with the parts of whorl C.

Usually, there is an accessory trace near the zone of
fusion between a segment of whorl C-D and one of whorl E.
This trace forks, one branch extending up into the broad
segment (whorl C-D) near its margin, the other branch
passing into the marginal portion of the staminode (whorl E).
It should be noted, also, that the accessory traces are
Joined to the staminode midribs at the base of the tube, as
shown in Figure 17.

In an attempt to determine the possible significance
of this accessory venation, one should refer to;Figures l9
and 20, which show the species of Manilkara, in which the
dorsal appendages (whorl D) are distinct from the segments
of whorl C- The reader may compare these Figures with
Figure 17, and picture the slightly lobed lower part of each
margin of a broad segment as representing the vestige of a
dorsal appendage. In this case, the accessory trace may be
comparable to a branch diverging from the midrib of the
staminode into an adJacent dorsal appendage. In some of the
broad segments (whorl C) in Figure 17, the branches of the
midrib diverge at a wide angle to pass into the small lobes
(the vestigial dorsal appendages) on either side; in one
broad segment, the angle of divergence of the right branch
of the midrib is so wide, that, as this branch forks, the
right arm passes into the adjacent staminode, while the left

one continues up near the lobed margin of the lower part of

33

the segment (see Figure 17). This trend.of wide divergence

of the branches of the midrib toward the slightly lobed margin
may possibly be interpreted as a vestige of the vascular con-
nection between a segment of whorl C and its adjacent dorsal
appendages, as found in the flowers represented by Figures 19
and 20. As shown in these Figures, one branch on each side
diverges at a wide angle toward the margin from.the midrib of

a segment of whorl C, to pass into the dorsal appendage.

These observations seem to indicate, that when whorls C,
D, E, and.F are all distinct and present, they are connected
by a prominent system of vascularization, and that a reduction
of this vascular pattern apparently takes place in prOportion

to the amount of fusion of whorls C and D.

It is interesting to note, that an interpretation of the
vascularization of a segment of whorl C-D as an androecium
segment (rather than as a corolla segment) seems to be in
accord with the telome theory of the phylogeny of stamens, as
expounded and illustrated by Wilson (1937, 1942). This theory
describes the plant body as one axis, branching dichotomously,
with some branches being;sterile, and others containing solitary
sporangia at their tips. Sepals and petals of modern flowers,
according to this theory, have arisen by the flattening, web-
bing, and.fusing of sterile telomes,vrhile the stamens and

carpels are the result of fusion of sporangiaebearing telomes.

According to the telome theory, the typical anther of four

34

pollen-sacs originated in this way (Wilson, 1942): "The
stalks of two pairs of sporangia became reduced, so that

the sporangia were almost sessile and were contiguous.

Each pair of sporangia then became folded back upon a
penultimate unit of the branch system and became fused with
this unit." The author demonstrates this theory of stamen
development in his paper by a series of illustrations, based
on the study of various types of stamens, taken from plants

of different families.

In some families with flowers having numerous stamens,
several of the stamens are supplied by branches from one
main trace, called the stamen fascicle trace, which arises
directly from the central cylinder. An example of this
condition.may be found in the family Annonaceae (Wilson,
1937), which is closely related to the Sapotaceae (Hutchin-
son, 1926). Note the similarity in the sapodilla flower:
The single strand departing from the stele appears to be
homolOgous to the stamen fascicle, since it branches at the
‘base of the tube to supply members of thevvhorls C, D, E,
and_F. In some cases, the filament trace has become fused
for its lower part with the midrib of the segment to which
it is attached. According to‘Wilson (1937), a fascicle
'trace may represent either lateral fusion of stamen traces
(in case a flower is a compressed shoot) or, by the telome
theory, the rachis of a primitive branch system with terminal
sporangia. The essentially dichotomous nature of vascular

'branching in the tube composed of whorls C, D, E, and F, in

35

the sapodilla and related flowers further implies the androecial

nature of this structure.

Upon morpholOgical examination of several related species

of Manilkara, Gilly (1943) found, that in some, the members of

 

whorl C (the so-called "petals" of authors) are of long and
narrow shape, while the dorsal appendages (whorl D) are broad
and petaloid. .An example of this type, Manilkara tabogaensis,
is illustrated in Figure 20. In E. emarginata (Figure 19) the
reverse is true. The shape of the staminodes (whorl E) also
varies in different species of Manilkara, ranging from short

and recurved to long and narrow to broad and petaloid.

The fact that the same range of variation in.shape and
size may be found in members of whorls C, D, and.E, as well
as the similarity of the nature of the vascular supply in these
whorls, add further support to the assumption that the tube

is an androecium, rather than a corolla.

Finally, it is noteworthy, that during the process of
fruit develOpment the so-called "corolla-tube" (whorls C, D,
E, and F) is lifted or torn out of the flower. The two outer
whorls (A and.B), on the other hand, are persistent at the
base of the ripe fruit.

SUMMARY

After experimenting with various methods of dehydration,
clearing, bleaching, and.ataining, it was found that the most
useful techniques in working with dried Sapodilla flowers from

36

herbarium specimens were dehydration in ethyl alcohol,
clearing in xylol, bleaching in peroxide, and.ataining
in Methyl Green and Bismarck Brown.

Although the venation in members of the two outer
floral whorls is similar, the source of vascular supply
to each is different. Two lateral veins, arising from a
higher level, in addition to the midrib, are present in
each sepal. These lateral sepal veins arise from the same
level as the midrib of the petals. The entire venation of

each petal originates from the single petal trace.

The traces passingginto the androecial tube diverge
from the main cylinder at a higher level than the petal
traces. The venation patterns of the exterior staminodes
(the "corolla-lobes" of authors) and of the inner staminodes

are very similar, and essentially dichotomous.

In the androecium of those forms of Manilkara zapotilla,
where the union between the exterior staminodes (the "corolla-
lobes" of authors) and the "dorsal appendages" apparently is
not entirely complete, an interesting type of accessory vena-
‘tion occurs. This may represent vestigial traces of the vas-
cular supply of the dorsal appendages. In some specimens
examined, the prominent accessory venation is replaced by a
network of small veinlets connecting the main veins of the

exterior and interior staminodes.

In specimens examined of two other species of Manilkara,

37

where the appendages are distinct from the exterior staminodes,
each “dorsal appendage" is supplied by traces from both the
exterior staminode to which it is attached, and the interior
staminode adjacent to it on the other side.

In one specimen of Manilkara emarginata, stamens with
three traces were found to be present, but lack of time pre-
vented further investigation. This is an interesting phenom-
enon, because it was observed in but one of the flowers examined

in this study.

The principa1.venation of the ovary consists of one dorsal
and one ventral trace per carpel, the dorsal bundles extending

up through.the style.
CONCLUSIONS

The anatomical study reported in this paper supports
morphological observations which indicate that the so-called
"biseriate calyx“ of authors actually consists of an outer
series of sepals and an inner series of petals. Furthermore,
the so-called "corolla-tube" appears to be an androecium.
This assumption conforms to the telome theory of the origin

of stamens;

The shape of the members of any of the sterile androecial
wherls may vary, in different species, from quite narrow to

very broad and petaloid.

The single traces, passing from the vascular cylinder of

38

the peduncle into the androecial tube and dividing to form
the midribs of stamens and.both the exterior and interior
staminodes, may be homologous to stamen-fascicle traces in

other plant families, such as the Annonaceae.

On the basis of such an interpretation of the structure
of the flower of the Sapodilla and related species of Magi;-
$222. the classical description of the family Sapotaceae
cannot be applied to the subfamily Mimusopoideae. This sub-
family would have to be described as polypetalous, with a
highly evolved androecial tube, separate from the corolla,
whereas the subfamily Sideroxyloideae is sympetalous, with
epipetalous stamens. The subfamily Madhucoideae, as yet not
investigated from an anatomical standpoint, appears to be

allied with the MimusOpoideae rather than the Sideroxyloideae.

 

‘!
E
1,.

Figure 1.

Figure 2.

Figure 3 .

39

PLATE I

Explanation of Figures

Floral diagram of Sapodilla, Manilkara zapotilla
(Jacq.) Gilly; the various whorls of the flower are

indicated by letters A—G (see text for explanation).

Floral diagram of a member of the Balata-complex,
Manilkara bidentata (A .DC.) Chev.; the various whorls
of the flower are indicated by letters A-G (see text

for explanation).

Diagrammatic representation of principal elements
of the vascular skeleton of a flower of Manilkara
zapotilla (drawn from Eyerdam 212); the flower is
represented as partially dissected; two sepals, one
petal, and one-sixth of the androecium are shown.
The letters A to G, inclusive, indicate levels at
which Figures 4-10 are drawn. Sepal midrib traces
are represented by solid black; lateral sepal traces
are stippled; petal traces are horizontally lined;

and an androecial trace is cross-hatched.

 

§

......-....J
.. __--—--_—..-___C
B

i r-’-----------

 

 

 

 

 

 

 

L
| I

 

 

 

4O

PLATE II

Explanation of Figures

Figures 4-10: Manilkara zapotilla, diagrammatic cross-sections

 

through a flower (drawn from Alexander s.n.); for'
detailed eXplanation, see text.
Figure 4. Cross—section from level A of Figure 3.
Figure 5. Cross-section from level B of Figure 3.
Figure 6. Cross—section from level of Figure 3.
Figure 7. Cross-section from level of Figure 3.

Figure 8. Cross-section from level of Figure 3.

“THUG

Figure 9. Cross-section from level of Figure 3.

of Figure 3.

Q)

Figure 10. Cross-section from level

PLATE II

 

 

Figure 11.

Fimue l2.

Figne 13.

Figure 14.

41
PLATE III

Explanation of Figures

Manilkara zapotilla, diagrammatic longitudinal
section of flower, (drawn from Broadway s.n.);
apices of style, anther, androecial tube, petals,
and sepals are not shown. Letters A-G, inclusive

indicate floral whorls as described in text.

Manilkara zapotilla, a single sepal (drawn from
Eyerdam 212) showing venation; A is midrib trace,
while B and B' are lateral traces (approximately
x 7.5).

Manilkara zapotilla, a single petal (drawn from
Eyerdam 212) showing venation (approximately
X76).

Manilkara zapotilla, inner view of one-third of
the androecial tube, showing pattern of principal

venation (drawn from Eyerdam 212; approximately
x705).

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

L
w

‘ ‘hh-uw‘dm I-I: menu “51/

Figures 15-20.

42

PLATE IV

Explanation of Figures

Drawings representing inner view of one-

third of the androecial tube, from flowers of the

Sapodilla and related species, showing venation

pattern (approximately x'7.5). One stamen and

one staminode in each drawing have been cut off to

expose the venation of the outer androecial seg—

ments.

Figure
Figure
Figure
Figure

Figure

Figure

15.
l6.
17.
18.
19.

20.

Manilkara
lganilkara
Manilkara
Manilkara

Manilkara

Manilkara

zapotilla (drawn from Lundell 599) .
zapotilla (drawn from £923.Z§§Q)°

zapotilla (drawn from Harris @223) .
zapotilla (drawn from Matuda 323) .

bahamensis (drawn from Seibert

1311) .

tabOgaensis (drawn fromflIoodson,
Allen, and.Seibert
IESEE.

P LATE EZ

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

20

43

BIBLIOGRAPHY

Arber, A. 1933. Floral anatomy and its morphological
interpretation. New PhytolOgist 32: 23l--242.

Baillon, H. 1892» Sapotacées. Hist. P1. 11: 255--304.

Cronquist, A. 1945. Studies in the Sapotaceae--Iv. The
North American species of Manilkara. Bull. Torrey
Club 72: 550-—562 .

------------ . 1946. Studies in the Sapotaceae--II. Survey
of the North American genera. Lloydia 9: 241-292.

Douglas, G. E. 1936. Studies in the vascular anatomy of
the Primulaceae. Am. Jour. Bot. 23: l99—-2l2.

Dubard, M. M. 1915. Les Sapotacées du groupe dos-Sider—
oxylinées-MimuBOpées. .Ann. mus. Col. Marseille III.
3: l—-62.

_Eames, A. J. 1931. The vascular anatomy of the flower*with
refutation of the theory of carpel polymorphism. Am.
Jour. Bot. 18: l#7--188.

----------- . and.L. H. MacDaniels. 1947. An Introduction.
to Plant Anatomy. Chapt. Xiii, pp. 34l--379. 2d ed.
McGrawAHill. New York and London.

Engler, A. 1903- Sapotaceae. MonOgr. Afrik. Pflanzenfam-
8: l--88.

Gilly, C. L. 1942. The nomenclature and systematic posi-
tion of the chicle and.Bapodilla trees. pp. 1--lO2.
fig. l-—14A. Unpublished.A.M. thesis in library of
Columbia University, New York.

44

Gilly, C. L. 1943. Studies in the Sapotaceae, II. The
Sapodilla--Nispero complex. Trap.'Woods 73: 1--20.

Hunt, K. W. 1937. A study of the style and stigma, with
reference to the nature of the carpel. Am. Jour.

Bot. 24: 288--295.

Hutchinson, J. 1926. The Families of Flowering Plants.

I. Dicotyledons. pp. 1--328. MacMillan and Co. Ltd.
London.

Johansen, D. A. 1940. Plant Microtechnique. 1st ed.
McGraw-Hill Co. ‘New York, London.

Lam, H. J. 1925. The Sapotaceae, Sarcospermaceae and
Boerlagellaceae of the Dutch East Indies and surround-
ing countries (Malay Peninsula and Philippine Islands).
Bull. Jard. Bot. BuitenZ. III 7: 1--269.

--------- . 1927. Further studies on Malayan Sapotaceae I.
Bull. Jard. Bot. Buitenz. III. 8: 381--493.

--------- . 1938. Monograph of the genus Nesoluma (Sapota-
ceae). Occas. Pap. BishOp Mus. 14: l27--l65.

--------- . 1939. On the system of the Sapotaceae, with
some remarks on taxonomical methods. Rec. Trav. Bot.
Neerl. 36: 509--525.

--------- . 1941. Note on the Sapotaceae-MimusOpoideae in
general and.the Far Eastern Manilkara allies in par-
ticular. Blumea 4: 323--356.

Lecomte, H. M. 1920. Une Sapotacée nouvelle du Congo.
Bull. mus. Hist. Nat. Paris 26: 5349539.

Lee, A. B. 1937. Microtomist's Vade-Mecum. P. Blakiston
and Son, Philadelphia .

45

McClung, C. E. 1937. Handbook of MicroscOpical Technique.
2d ed. Paul B. Hoeber, Inc. New York.

Rawlins, T. E. 1933. PhytOpathological and Botanical
Research Methods. John Wiley and Sons, New York.

Swamy, G. Lo 1949. Further contributions to the morpholOgy
of the Degeneriaceae. Jour. Arnold Arb. 30: l--37.
Tillson, A. H. 1940. The floral anatomy of the Kalanchoideae.

Am. Jour. Bot. 27: 595--600.
Wilson, C. L. 1937. The phylogeny of the stamen. Am. Jour.
. Bot. 24: 686-699.
------------ . 1942. The telome theory and the origin of
I the stamen. Am. Jour. Bot. 29: 759--764.¢
‘Woodcock, E. F. 1945. Laboratory Manual for Plant Histology.
mimeozraphed. Michigan State College; East Lansing,
Michigan.

 

 

 

 

ISE ONLk’xg‘G’m USE om.

 

 

I w‘. ‘ YI_ "
, ,i..‘\'. ‘l

J
I
I
c
I
I f
. l.
u 1." '
J r
.7. '-
| ' - " I
., . I».
. I!
| \
- n .
I
0 ‘ I
?
, Y
.' |- A
‘ g n
I I
' 'I
. I"! f
l I
‘ I
\ a l. ‘
)‘~
. ‘ yr "9‘ ‘-
I I, ”
~ . . .
\ _K' 1‘ "
.h
v
'.w. b
. n '
l' ‘
t .
.' ). ’ x'
, .. ‘ I _"
. Q
- (VJ,
‘ .... l t ‘\ F.
n: he "u

vi;

MIC CHIGAN STATE UNIVERSITY LIBRAR

II 3II||13131|II3|III