_

/<?/<2 ->3Z>

Vol. V 1919 No. 1

CONTRIBUTIONS

FROM THE

Botanical Laboratory

OF THE

University of Pennsylvania

University of Pennsylvania
Philadelphia

1919

CONTENTS OF VOLUME V, NO. 1.

Page

1 . The Macroscopic and Microscopic Structure of some Hybrid

Sarracenias Compared with that of their Parents. By

Alice Mary Russell, B.S., M.S. (With plates i, ii, iii, iv, v.) 3

2. A Comparative Study of the Structure and Saphrophytism

of the Pyrolaceae and Monotropaceae with Reference
to their Derivation from the Ericaceae. By Margaret
\Y. Henderson, B.S., M. A 42

Vol. V 1919 No. 1

1-iBRARY

ni;w YORK

BOTANICAL

GARDEN

CONTRIBUTIONS

FROM THE

Botanical Laboratory

OF THE

University of Pennsylvania

University of Pennsylvania

Philadelphia

1919

Y.5*

The Macroscopic and Microscopic Structure of

some Hybrid Sarracenias Compared with

that of their Parents

By Alice Mary Russell, B.S., M.S.

[Thesis presented to the Faculty of the Graduate School in partial fulfillment of the
Requirements for the Degree of Doctor of Philosophy.]

CONTENTS

Historical Review 4

Natural Distribution of Species and Hybrids Selected for

Study 8

Comparison of Parents and Hybrids

S. purpurea, S. flava, S. Catesbaei 9

S. flava, S. Drummondii, S. Moorei 18

S. Sledgei, S. Drummondii, S. areolata 24

Comparison of Flowers 27

Description of Structure of Glands 32

Ovarian Gland Structure 33

Conclusions 35

Bibliography 39

Description of Plates 4°

3

Russell— Comparison of the Structure of Hybrid

Historical Review

Before Tournefort had named the genus Sarracena (i) and
Linnaeus (2) had accepted the name, the group of the Sarra-
cenias was already known to the early settlers in North America.
They collected the plants and sent them to Europe as inter-
esting exotics, where they were carefully described in botanical
publications. As early as 1570, Lobel described specimens of
the group which had come to his attention. Clusius (1601) (3)
figures 5. purpurea and Parkinson (4) copies his figure and
adds a note which seems to indicate that he knew S. flava as
well. Concerning this plant, which he calls "The Hollow
Leaved Strange Plant of Clusius," he writes: "This strange
plant hath such strange leaves, as the like are seldome seene
in any other that we know growing, for they are nine or ten
or more, rising from the head of a small long roote, each by
itselfe, being small below, and growing greater upward, with
a belly as it were bunching forth, and a bowing backe, hollow
at the upper end, with a peece thereon like a flappe, and like
unto the flower of Aristolochia, or Birthwort, and round at
the mouth like a halfe circle, full of great darke purplish veins
on the inside; the whole leaf is of thicke substance almost like
unto leather; among these leaves sprang a stalke but was broken
short off, so that what flower or seed it bore could not be ob-
served. This was sent to Clusius from Paris by one that re-
ceived it from Lishbone in the same manner. But of late
Master John Tradescant the younger found this very plant
in Virginia, having his toppe thereon, which he brought home
and groweth with him, which I here show you with Clusius
his figure. The leaves are longer, narrower and not bellying
out, and the flower is borne at the top of the roundish seed
vessell." The specimen sent by Tradescant was probably 5.
flava.

Plukenet (6) (7) was familiar with both S. flava and 5. pur-
purea, since he gives very accurate figures of both species.

John Ray (8) gave a Latin translation of Parkinson's descrip-
tion already quoted. The actual specimen described by him,
however, was a natural hybrid between S. flava and 5. pur-
purea, and was the first natural hybrid collected. (See below.)

-J

Sarracenias with that of Their Parents 5

Tournefort named the genus Sarracena, in honor of Dr.
Sarrasin of Quebec (1), and described one species, S. canadensis.
Linnaeus accepted the genus name and described the two spe-
cies long recognized, S. flava and 5. purpurea.

Walter, in 1788 (9), described and named two new species,
S. minor and 5. rubra. S. psittacina was added to the genus
by Michaux in 1803 (11). Croom (12) described S. Drum-
mondii in 1835.

Since the above time, only one new species has been added,
5. Sledgei in 1906, by Macfarlane (17).

A few of the botanical publications of this time review the
genus as varying in composition:

"Flore des Serres" (13) gives seven species: S. flava (L.), S.
purpurea (L.), S. variolaris (= minor Walt. = adunca Smith),
S. undulata (Den.), 5. Drummondii (Cro.), S. rubra (Walt.),
S. psittacina (Michx.) (= calceolata = pulchella (Croom)). Chap-
man (14) gives all of the above except 5. undulata, which he
considers synonymous with 5. Drummondii. Hooker (15) men-
tions eight species but does not enumerate them. Boulger
(16), in reviewing the genus, has the six species: S. purpurea,
S. flava, S. rubra, S. Drummondii, S. psittacina, S. variolaris.

During the latter half of the 18th century, Sarracenias were
widely cultivated in European gardens. New varieties were
eagerly sought for exhibition, many new forms were introduced
from America and several artificial hybrids were produced.
Since each exhibitor appended a name to his own product, a
great confusion of names had arisen and a survey of the forms
under cultivation became most necessary. Dr. Masters, there-
fore, undertook the review in three numbers of the Gardeners'
Chronicle for the year 1881 (18). Here he gives a key to the
forms raised in English gardens and gives for each a short de-
scription from living specimens furnished to him. The forms
and species described by him are as follows:

1. S. psittacina (Michx.) A. D. C. Prod. XVII, p. 4.

2. 5. purpurea (L.) A. D. C. Prod. XVII, p. 4-

3. S. Chelsoni X (Hort. Veitch, G. C. vol. 9. P- " (rubra X purpurea).

4. 5. variolaris (Michx.) Croom. A. D. C. Prod. XVII, p. 6.

5. 5. Drummondii (Croom) A. D. C. Prod. XVII, p. 5 (var. alba G. C.

vol. 10, p. 281).

6. 5. undulata (Den.) = 5. Drummondii (Croom) Rev. Hort. i, p. 126.

Flore des Serres 7, A. D. C. Prod. XVII, p. 5- Index Amer. Bot.
p. 40.

6 Russell— Comparison of the Structure of Hybrid

7. S. rubra (Walt.) Flora Car. p. 152. (Croom) A. D. C. Prod. XVII, p. 4.

var. acuminata (A. D. C. loc. sit.) var. Sweetii (A. D. C.) Wat.
Index p. 40 = 5. minor (Sweet) = 5. rubra (Planchon).

8. 5. flava (L.) Sp. PI. Ed. i, p. 510. A. D. C. Prod. XVII, p. 5. var.

Catesbaei (Ell.) Bot. S. Car. = S. flava var. picta Hort. Bull. = 5.
Fildesii Hort. Williams.

Var. omata (Hort. Bull.).

Var. Rugelii (Shuttleworth) = erythropus Hort. Bull., A. D.
C. Prod. XVI, p. 6.

Var. limbata (Hort Bull.).

Var. maxima (Hort. Angl.).

Var. cristata (Hort. Bull.).

Var. atrosanguinea (Hort. Bull.).

Var. minima (Hort. Angl.).

9. S. Moorei X (G. C. 1874, p. 702 = 5. Drummondii X S. flava).

10. 5. Stevensii X (G. C. 1874, p. 738) = 5. flava X 5. purpurea).

11. 5. Williamsii X (Hort. Williams above 10).

12. S. Popei X (S. flava X 5. rubra).

13. 5. melanorhoda X (Hort. Veitch. 5. purpurea X S. Stevensii).

14. 5. formosa X (Hort. Veitch. 5. psittacina X 5. variolaris).

The above review constitutes the basis for the article on
Sarracenias in Nicholson's Dictionary of Gardening (20).

The history of the hybrids grown during this period is inter-
esting. In 1874 the first artificial hybrid was produced and
exhibited by Dr. Moore at the International Botanical Congress
in Florence. An abstract of Dr. Moore's paper upon the pre-
sentation of the hybrid is given in the Gardeners' Chronicle
1874, p. 738. Of the plant, which had S. flava as the female
and S. Drummondii as the male parent, he writes: "The plant
is as nearly as possible intermediate between those two noble
species of the genus, and no hybrid which has hitherto come
under my notice proves more decidedly than it does the marked
influence of the pollen of one plant applied to the stigma of
another. . . . During the months of April and May most of
the species flower and produce young leaves, after perfecting
which the plants rest six weeks or more, when some of the kinds
produce a second crop of leaves which remain fresh during
the winter and are more beautiful than those of the first crop.
This is especially the case with S. Drummondii. . . . S. flava
does not make a second growth of leaves in so marked a manner,
but rather inclines to rest during the winter months." "Now
it is in the mixture of the leaves that the intermediate state
of the hybrid is so strikingly exemplified. It makes a second

Sarracenias with that of Their Parents 7

growth of winter pitchers similar to 6". Drummondii, and these
are nearly as highly colored, but they decay much sooner than
those of the parent species and thus resemble more those of
the female parent, 5. flava. Further, the large stature of the
plant, nearly two feet high, the purple color of the flowers,
and, indeed, everything connected with it, shows that it holds
an exactly intermediate rank between the parents." Shortly
afterwards Messrs. Veitch and Sons were awarded a prize for a
hybrid between 5. flava (female) and S. purpurea (male), and
named in honor of the gardener whose product it was, 5. Stevensii.
A hybrid of the same parentage, but of natural origin, was
exhibited under the name S. Williamsii, in honor of the man
who discovered it among a mass of 5. flava shipped to him from
the Southern States (22). Other artificial hybrids were shortly
afterwards exhibited and are noted above by Masters.

The most recent review of the genus is that of Macfarlane
in "Pflanzenreich" (23). All of the species and hybrids noted
above are given, but a new species is here added, 5. Sledgei (17).
Macfarlane reviewed and fully described the supposed species
6". Catesbaei (24) in order to eliminate the confusion which had
arisen through ascribing to it forms such as S. rubra, S. flava,
etc., because of Elliot's meager description, and the misplacing
of his type specimen. The plants used as type specimens by
Macfarlane were first obtained through Dr. Sledge, of Mobile,
who sent them north, where they were grown in the Sarracenia
House at the University of Pennsylvania Botanical Garden.
The pitchers and flowers were typical, Macfarlane found, of
plants cultivated abroad under the names S. flava, var. cristata,
S. flava, var. picta, or 5. flava, var. Catesbaei. A visit to the
region established the fact that these plants grew in pure stands
and were indigenous to the locality, so they were used as type
forms on which the description of the supposed species S. Cates-
baei was based. Later the type of Elliot for the above species
was found in the Charleston Museum, and proved to be a
specimen of the natural hybrid between 5. flava and S. purpurea.
A new name was therefore applied to the genus described by
Macfarlane, 5. Sledgei. In regard to the new species he writes:
"So far as accurate records show, 5. Sledgei seems to be con-
fined to the Gulf region between the Alabama River and Eastern
Texas, over which area it may at times be extremely abundant.

8 Russell— Comparison of the Structure of Hybrid

It is here only associated with S. psittacina and 5. purpurea,
both of which flower from one to two weeks later. In spite
of this there is every likelihood that hybrids will in time be
reported, for the flowers of all species last for 14 to 21 days" (17).

Since the time this was written, the author himself has found
S. Drummondii growing with 5. Sledgei, and with them quan-
tities of the cross which he has named S. areolata. Macfarlane
(26) gives the following note concerning this plant:

"It is a frequent hybrid wherever both parents are present
near each other. It is especially abundant near Mobile, Ala.,
and westward for thirty miles."

Natural Distribution of Species and Hybrids Selected

for Study

Of the forms to be studied, S. purpurea is the best known
and most widely distributed (23). It is found from Labrador
through Newfoundland, Quebec, Manitoba, Michigan, Wiscon-
sin, and from thence southward to Indiana. Along the coast
it extends from New England to Florida in the swampy regions
back as far as the Alleghany Mountains. In the south it is
plentiful about the Gulf region from Florida to Louisiana.

5. flava has a more restricted distribution (23). It is more
or less abundant from southern Virginia, through North and
South Carolina, Georgia, Florida, and eastern Alabama.

Therefore, the hybrid between these species might be found
in the region from North Carolina to eastern Alabama. It is
reported from Wilmington, N. C, and Ponce de Leon, Florida.
Possibly additional localities may be reported from Georgia (23)
and South Carolina.

S. Drummondii is found only between south central Georgia
and western Alabama (23). Since 5. flava is never found west
of the Alabama River, the hybrid would not be expected out-
side of southern Georgia, Florida, and eastern Alabama. Speci-
mens have been gathered in Georgia (Americus), in Florida
(Milligan and Crestview), and from Alabama (Bay Minette
and Deer Park).

S. Sledgei, most recently described species, is limited to the
region about the Gulf, west of the Alabama River (23). There
are specimens preserved from Alabama, Louisiana, and east-

Sarracenias with that of Their Parents 9

ern Texas. Additional localities may yet be reported. The
hybrid, 5. areolata, with 5. Drummondii, is therefore to be
sought only between Alabama River and western Alabama.
It is especially plentiful at two points, Theodore, Alabama,
where it was first found, and Deer Park.

The above are the species and hybrids selected for study.
The material used was from natural sources indicated above,
transplanted to the University of Pennsylvania Botanical
Garden.

It might be well to note here the other natural hybrids re-
ported by various collectors. Besides the above three, there

are given (23; 17):
5. flava X S. minor.

S. minor X 5. psittacina (grown in University of Pennsylvania).
5. rubra X 5. Drummondii.
S. psittacina X S. purpurea.

S. rubra X S. purpurea (grown in University of Pennsylvania).
5. Drummondii X S. purpurea (grown in University of Pennsylvania).
5. Sledgei X 5. purpurea (grown in University of Pennsylvania).

The study was limited to three groups, with one parent com-
mon to two hybrids in order the better to ascertain the action
of a given parental characteristic.

Comparison of Parents and Hybrids

Set 1. S. purpurea, S. flava, and S. Catesbaei (S. Stevensii)

A. Naked Eye Characters

The general habit of 5. purpurea is shown well in PI. I, fig. i,
with its inflated pitchers, and the decumbent leaves. S. flava
presents a striking contrast to S. purpurea (PI. I, fig. 2). The
pitchers are erect, slender, and gradually expanded upwards.
In the hybrid (PI. I, fig. 3) the pitchers lean at an angle of 450,
and are slightly inflated in their middle portion.

In height the three forms present a wide range. S. purpurea,
the lowest, averages 12-15 cm., though specimens can be
found 35 cm. in length. S. flava may be from 20-100 cm. high,
but average pitchers measure 60-70 cm. from base to the tip
of the lid. The hybrid averages 24-36 cm. in length.

The "wing," or fused laminar faces in front of the pitcher
proper, is wide in S. purpurea (PI. I, fig. 1) and tends to have
an undulating margin. In S. flava the wing is very narrow

io Russell— Comparison of the Structure of Hybrid

(PI. I, fig. 2) and extends the whole pitcher length. In S. Cates-
baei the wing is wider below, half as wide as in S. purpurea,
and of much the same shape (PI. I, fig. 3).

In 5. purpurea the inflated pitcher is constricted below the
rim around the mouth (PI. I, fig. 1). In 5. flava the pitcher
has no suggestion of a constriction, but rather expands widely
at the rim. In S. Catesbaei the pitcher is slightly constricted
below the rim (PI. I, fig. 1), although not so strongly as in 5.
purpurea.

The color of the pitchers of S. purpurea is: base green with
a reddish network of veins over the outer pitcher surface, par-
ticularly above. The lid is especially strongly marked with
crimson-purple reticulations. (Variations: all green; with
graded transitions to crimson-purple suffused over the surface
almost uniformly.) 5. flava, in the most common type, has
a greenish yellow pitcher and lid, the latter with a deep crim-
son patch at its base. The marking is usually band shaped,
about 1-2.5 cm. wide. (Variations: purplish veining on upper
pitcher and lid; bright green pitcher; or entire pitcher and lid
rich purple.) S. Catesbaei shows a variety in coloration, accord-
ing to the variations noted above in the parents. An average
type represented in fig. 3 has a green pitcher with red-purple
veinings, less heavily developed than in S. purpurea. The red
banding in the throat of the S. flava parent is reproduced, but
more dilute in color. (Variations: uniformly green, more pro-
nounced purple markings, or entirely purplish.)

The lid shape in the three forms exhibits a nearly balanced
relation. In S. purpurea the lid is reniform, with an undulate
margin (PI. I, fig. 1). The lower lateral portions of the lid
are prolonged into blunt lobes, which are bent forward about
the mouth of the pitcher, so that it is open only from the front.
Instead of overhanging the pitcher opening, the lid in this spe-
cies is somewhat curved outwards. In S. flava, the lid is ovate-
cordate, but prolonged in its median portion into a tip process.
The sides of the lid are not lobed as deeply as in 5. purpurea,
but the slight auricles are here bent sharply backwards, expos-
ing the throat of the pitcher. The lid slightly incurves above
the orifice. S. Catesbaei (fig. 3) is very nearly intermediate
in parental characteristics. The lid is more rounded in out-
line than S. flava, but possesses a slight tip process, not as pro-

Sarracenias with that of Their Parents n

nounced as in 5. flava. The lobes are intermediate in size,
in position also, since they stand straight out from the side of
the pitcher. The margin of the lid is wavy, with looser undu-
lations than in S. purpurea. The lid does not overhang the
pitcher mouth, but is erect.

B. Microscopic Study of the Pitchers
I. Methods.

Three methods were used in preparing the epidermal sur-
faces described below.

(a) Scrapings of fresh material were made, for comparison
with treated material to gauge possible shrinkage, or seeming
abnormalities.

(b) Scrapings of material macerated in KOH.

(c) Strips of pitcher parts were boiled in 25%~50% HN03,
to which a pinch of KCIO3 was added. When the material
appeared white, or bleached, it was quickly placed in water.
The mesophyl and subepidermal cells could be easily brushed
off and both upper and lower epidermis could be mounted
in acetic acid side by side for comparison ; sealed with asphaltum.
The nitric acid was not used in such strength that cell walls
were affected. The delicate hair striations appear clearly in
the photographs (PI. IV, figs. 15 and 16) made from such slides
with the aid of the Edinger apparatus.

Ordinary paraffin material, fixed in chrom-osmo-acetic or
weak chrom-acetic, was used for sections mentioned.

In counts given for a "field," it will be understood, unless
distinctly stated otherwise, that the standard "field" is that
of a No. 3 Bausch and Lomb eyepiece, and a No. 4 mm. Bausch
and Lomb objective.

Averages are based on 100 or more counts, and measurements
on 200 or more. Corresponding portions of corresponding
surfaces were carefully selected for comparison.

Outer Epidermis of the Lid

In all three forms, the outer epidermis of the lid shows cells
that are roughly quadrangular, with wavy walls. Normal
stomata are present, and simple unicellular hairs. In general
these hairs point upwards toward the edge of the lid. Glands,
similar to those of the interior pitcher area, are scattered over
this outer surface somewhat irregularly.

12 Russell— Comparison of the Structure of Hybrid

In S.flava the epidermal cells are longer than broad (PI. Ill,
fig. 9) with an angular wavy wall. Stomata are fairly numer-
ous, seven to a field, and of the same size as within the pitcher
(•035X.030 mm.). Glands are less numerous than within on
the specially modified surfaces, one only to a field averagely.
The unicellular hairs mentioned above are in 5. flava blunt,
with a decided bending or hooking in some, while in others
the axis is straight or merely slightly curved. They vary
greatly in size, from mere knobs .01 mm. long to .1, .2, .3 mm.
in length. PI. Ill, fig. 9 will show the range of size and shape
in these hairs. It will be noted that in all the surface of the
hair is raised in wart-like thickenings, which at times lie in
lines parallel to the axis of the hair, and suggest the rough be-
ginnings of the beautifully regular striae found on the hairs
of the inner pitcher surfaces. There are three hairs to a field.

The epidermal cells of S. purpurea (PI. Ill, fig. 8) are more
rounded wavy in outline than in S. flava. The stomata are
more numerous, 16-17 appearing in a field. The glands are
but slightly more frequent than above, 1.5 to a field. The
hairs are straight, and vary in size even more than in S. flava
.1, .2, .3, .4, etc., to .5-. 6 mm. They are not quite so numerous
as the hairs of S. flava, only two to an average field.

5. Catesbaei (PI. Ill, fig. 10) shows epidermal cells more
closely resembling S. purpurea. The stomata are fairly inter-
mediate in number — 11 per field. There is one gland per field
about as in the parents. The hairs, of course, inherit the
tendency to variability. Some hairs have the marked S. flava
tendency to bending, though not so strong. On the whole,
the range of size is much as in S.flava from .1 to .3 mm. There
is one hair per field, less than in either parent.

Inner Lid Region

The epidermal cells of the inner portion of the lid are in all
cases irregular in shape. In 5. purpurea they are about as long
as broad (.065 x .05 mm.) with very wavy wall.

In 5. flava the hairs are numerous, and compress the epi-
dermal cells without hairs into oblong cells (.05X.03 mm.)
with walls but slightly wavy. 5. Catesbaei shows cells resem-
bling S. purpurea in size and shape (.06 X. 047 mm).

Sarracenias with that of Their Parents 13

Stomata are frequent on this surface. In these three forms
they are of the same size, .035X.030 mm. In number they
are noteworthy, 5. flava having seven to a field, S. purpurea
three, and the hybrid shows an intermediate number — five.

Glands are numerous over this, the alluring, surface. In
5. flava there are 3-4 glands per field, in 5. purpurea 2-3. The
hybrid averages less than three.

This region is beset with more or less numerous long, stiff
hairs, directed downward. These prevent the insects from
pursuing any path other than that to the treacherously smooth
conducting surface. The hairs are thickened in parallel ridges,
varying in number according to the species.

In 5. flava (PI. IV, fig. 16) the hairs seem to be remarkably
uniform in length. Measurements show that over 50% are
.22 mm. long, while the remainder are .11 or .3 mm. in length.
They are very numerous — 10-11 to an average field.

5. purpurea (PI. IV, fig. 15) has hairs many times longer
than in S. flava. Besides these, there are shorter hairs, but none
as short as the longest hairs of S. flava. They vary — about
50% are 1.3 mm. long, 20% are .0-.9 mm. in length, and the
remainder are 1.8 mm. long or measure .9-1.0 mm. They are
quite scattered, one hair base appearing in two fields.

5. Catesbaei (PI. IV, fig. 17) has a variety of hairs, as has S.
Purpurea. Some are short, resembling those of 5. flava, .2 m.
long, while others approach the 5. purpurea type — 1 mm. long.
Over 50% are less than .4 mm. in length; 20% are .5-6 mm.
long; the rest measure fom .7 to 1 mm. Hairs of the length of
both parents are represented, since the shorter hairs of S. pur-
purea are .9-1 mm. in length, but by far the larger number of
hairs are closer to the S. flava type. There are but two hairs per
field.

The lower part of the tube, representing the conducting
surfaces in S. flava and S. Catesbaei, and the glandular surface
of 5. purpurea, has an exterior epidermis of cells of the same
character as those of the exterior of the lid. Hairs of the same
character are present, but more infrequent than on the lid.
These hairs point upwards, generally, toward the pitcher mouth,
and become more numerous on the upper portion of the tube.

Stomata are very numerous over this surface. In S. flava
and S. purpurea there are 14-15 stomata to a field; in the hybrid

14 Russell— Comparison of the Structure of Hybrid

fewer — io-ii. The glands present have a tendency to be
distributed along the strong ridges above the veins, especially
in 5. flava. Averages of counts from the surface between the
veins run about .5 gland per field, while on the veins 2-3 glands
appear in a field. In 5. purpurea, where the veins are not so
emphasized, the glands are equally distributed over the outer
surface, and there are two to a field. In 5. Catesbaei there
are fewer glands than in S. flava — one to a field, though here
there is a slight tendency for the glands to be more numerous
over the veins.

The ridges spoken of above are formed by the reinforcement
of the larger bundles on their inner and outer faces with a deep
sclerenchymatous development. Above such bundles the sub-
epidermal and epidermal layers become thickened enormously,
so that the whole thickened area stands out above the sur-
rounding epidermis. The ridges are very marked in 5. flava,
are not shown in S. purpurea, and are fairly conspicuous in
5. Catesbaei.

The epidermis above the bundles undergoes a peculiar trans-
formation. In S. purpurea, the wavy walled epidermal cells
become elongated above the bundles. In 5. flava, the epi-
dermal cells are elongated, straight walled cells, much thick-
ened. S. Catesbaei has these cells elongated slightly, heavily
thickened — in fact about intermediate in character. Along
these lines of cells there are no stomata, nor hairs; therefore
they form the easiest paths which the insects can pursue up-
ward toward the rim of the pitcher.

Conducting Surface

Below the lid surface in all forms, the epidermal cells become
gradually polygonal in shape, and the hair processes become
shorter and sharper. Every cell becomes prolonged into a
pointed projection directed downward. The cells thus form
a scale-like slippery covering for the inner pitcher surface to
a varying depth. In 5. purpurea it forms a narrow band, 1-2
cm. wide; while in S. flava and 5. Catesbaei it extends over at
least one-half of the pitcher depth.

The conducting cells of these three forms are interesting
in length relation. In 5. purpurea the tip process is but little
more than a knob (PI. V, fig. 22) and the cell measures .05 mm.

Sarracenias with that of Their Parents 15

to the end of the projection. S.flava has a long, fine tip (PI. V,
fig. 23) and the cell measures .07 mm. in length. The upper
portion of the conducting surface of S. Catesbaei shows an ex-
actly intermediate size .06 mm. (PI. V, fig. 24). These cells
are all beautifully thickened with striae, as are the lid hairs.

L. S. Pitcher Rim

A longitudinal section through the rim of the pitcher men-
tioned above shows an interesting relation. In S. purpurea
the rim is rolled outward and under, in two turns. In S. flava
the rim is rolled over once, rather loosely. 5. Catesbaei is
rolled over, and slightly curved around again, or is rolled once
and half.

The tip region in S. purpurea has three layers of thickened
cells, while in S. flava the whole tip region is thickened. 5.
Catesbaei shows about an intermediate amount of thickening.

The conducting cells, with their fine processes, do not begin
to show in S. flava at the tip, but considerably below the tip
on the outwardly rolling portion.

Below the conducting cells in 5. purpurea there are two layers
of thickened cells, closely united to each other and to the epi-
dermis. In S.flava there is but one such layer. In S. Catesbaei
there is one regular continuous layer adjoining the epidermis,
and a second discontinuous layer below it. The mesophyl in
5. purpurea is loose, with large thin-walled cells. S. flava has
a more compact and more shallow tissue than in 5. purpurea.
The mesophyl in 5. Catesbaei is not so loose as that of S. pur-
purea, nor so compact as that of S. flava.

The outer epidermis in all three forms is composed of thin-
walled cells. The stomata have their guard cells raised above
the level of the epidermis in all three.

Below the outer epidermis is a region of regular large, thin-
walled cells closely applied to each other. These cells form
a false palisade tissue in the pitchers. In 5. flava there are
3-4 layers of these cells; in S. purpurea 2-3 layers; in S. Cates-
baei there are 2 layers. These cells contain numerous chloro-
plasts, and form a continuous layer, except where interrupted
by stomatal chambers.

In S. purpurea, below the narrow conducting surface there
is a wide expanse of glandular surface, reaching down to the

16 Russell — Comparison of the Structure of Hybrid

detentive. This region, absent in S. flava, is characterized
by large regular wavy- walled cells with numerous glands —
6.6 per field. These cells are smooth, without a suggestion
of any hair process. The lower conducting area in the hybrid
represents a crossing of the lower conducting area of 5". flava
with the glandular area of S. purpurea. There is no marked
difference between the upper and lower portions of the con-
ducting surfaces in the hybrid, excepting for a tendency toward
lengthening of the cell process, and for the presence of numerous
glands.

On the upper conducting regions of 5. flava there are 3 glands
per field; in S. purpurea there are 4 per field; in S. Catesbaei
2.9 per field in this region. The lower conducting surface in
S. flava is devoid of glands; the glandular surface of S. purpurea,
as has been mentioned, has 6 glands per field; in the hybrid
there are 3, about half as many as are on the glandular area in
5. purpurea.

The lengthening of the cell process noted above for the hybrid
in this region may be due to the influence of the S. flava type
of conducting cells, or merely be the exhibition of the tendency
shown in all forms for the conducting cell process to become
prolonged deep in the pitcher.

In the extreme lower conducting and upper detentive sur-
face, the hybrid presents an interesting condition. This region,
which represents a crossing of glandular-detentive area of S.
purpurea with conducting-detentive of 5. flava, exhibits a most
confused zone about 1-1.5 cm. wide. There are several types
of cells shown in this region:

First, normal conducting cells which become gradually elong-
ated and give place to detentive surface cells.

Second, cells intermediate in size and shape between the
glandular surface cells of S. purpurea, and the detentive sur-
face cells of S. flava. These are long narrow cells with wavy
walls. They appear at first isolated as islands of two or more
cells, in the lower conducting surface; becoming more numerous
below, and gradually giving place to normal detentive-surface
cells. These cells represent a blending of the glandular sur-
face cells of 5. purpurea, with the unmodified epidermal cells
of the detentive surface of S. flava.

Third, normal detentive surface cells, polygonal in shape,
some bearing long hair processes. These appear isolated in

Sarracenias with that of Their Parents 17

the region where the blended detentive-glandular cells are most
numerous.

Fourth, normal glandular cells of S. purpurea. These cells
are reproduced in patches, isolated, in the region of the detentive-
glandular cells.

Fifth, normal detentive epidermal cells with or without hair
processes, intermediate between the types of 5\ flava and 5.
purpurea.

Sixth, normal glands; two per field. These occur down to the
purely detentive zone, where they cease.

A similar case, such as above, is presented in a hybrid be-
tween S. minor, without a glandular area, and S. purpurea,
and has been discussed shortly by Macfarlane (28). He con-
siders that this region represents a modification of the upper
detentive surface. In both hybrids it seems to be shown that
the conducting-surface cells of one parent are incapable of
blending with glandular surface cells of S. purpurea, since no
wavy-walled cell with a prolongation characteristic of the con-
ducting surface type ever appears. In other words, this rep-
resents a clear case of the so-called bi-sexual hybrid (29), or
of particulate inheritance (38).

The detentive surface presents no such confused condition,
rather a perfect blending, cell for cell. In this region, in these
forms, there are no stoma ta nor glands present. In all, the
epidermal cells are polygonal and contain a fair amount of
tannin (30). The hair cells are remarkable in relation. In
S. purpurea they are long — 1.5 mm. — and rather infrequent —
.7 to a field. Those of S. flava are quite short — .37 mm. —
but very numerous — 5.5 per field. In 5. Catesbaei the hairs
are intermediate in size and number, for they are .88 mm. long,
with 2.8 per field.

Lower Pitcher Region

T. S. of pitcher at lowest part of tube shows in all a hollow
cylindrical portion, representing the T. S. of the pitcher proper.
From the front extends the more or less wide wing, representing
the fused laminae (23). In the cylinder the bundles are dis-
posed at intervals, larger and smaller alternating, with the
xylem on the inner side, toward the pitcher cavity. In the
wing, two rows of bundles face each other, xylem opposed to

1 8 Russell—Comparison of the Structure of Hybrid

xylem. The things especially to be noted in comparing parents
and hybrids are: the number of the subepidermal layers of
cells, the depth and character of the mesophyl tissue, the rela-
tive amount of sclerenchymatous tissue developed on the
inner and outer side of the bundles, the number of subdetentive

layers.

In all three forms, the outer epidermal cells are regular with
heavily cutinized walls. The stomata have their guard cells
slightly raised above the epidermal level.

Below this are the subepidermal layers of cells rich in chloro-
plasts, forming on the morphologically lower surface a false
palisade. In S. purpurea there are 3-4 layers of cells, large
and thin-walled. In S. flava there are 4-5 layers of these cells.
They are smaller and more compactly arranged. 5. Catesbaei
resembles more S. purpurea in that there are three subepidermal
layers, which have larger cells less compactly arranged than in
S. flava.

The mesophyl is in 6". purpurea deep and loose. The cells
are large, and intercellular spaces are frequent. S. flava has
a very narrow mesophyl zone. The cells are regularly arranged
with few intercellular spaces. In S. Catesbaei the mesophyl
is deep and spongy as in S. purpurea. In the mesophyl the
bundles are distributed. In 5. purpurea they are not strongly
reinforced with sclerenchyma tissue. In 5. flava the rein-
forcement is pronounced, and involves the subepidermal and
epidermal layers as well in the larger bundles. In 5. Catesbaei
the development is about intermediate between the parents.

There are two or more layers of cells below the detentive
surface which are thickened together with the epidermis. It
is this which forms the so-called "absorption zone" mentioned
by Fenner (31) and others. The contents of these cells are
either gathered into large masses, or broken up into finely
granular substance. These layers give a strong positive reac-
tion for tannin. In 5. purpurea there are 1-2 layers of sub-
detentive cells. In S. flava there are two layers. In 5. Cates-
baei there are three layers.

Set 2. S. flava, S. Drummondii and S. Moorei

The second hybrid of the series, between 5. flava and S.
Drummondii, was the first artificial hybrid between species
of Sarracenias to be exhibited. Though it was named in honor

Sarracenias with that of Their Parents 19

of its exhibitor, Dr. David Moore (21), it was also widely known
as S. Mandaiana, which name was applied to the natural hy-
brid.

Both parents are strong and beautiful types, and as one
might expect, as already noted in Dr. Moore's paper, quoted
above, the hybrid is also a striking form. It is intermediate
in appearance between the parents where the parental char-
acteristics are capable of blending; but exhibits an apparent
marked divergence toward one parent where a single char-
acteristic can be inherited. This is particularly noteworthy
in the pale areas on the lid and tube of 5. Drummondii.

S. Drummondii (PI. I, fig. 4) is, like 5. flava (PI. I, fig. 2),
tall, strong, erect, with pitchers ranging from 30-90 cm. high.
Average specimens usually measure 60-70 cm. 5. flava, it will
be recalled, is about of the same height. S. Moorei (PI. II,
fig. 5) is about 65-70 cm. high.

In both parents and hybrid the pitchers are slender below,
increasing gradually in width to the flaring rim. The pitchers
of 5. flava seem on the whole to be more robust than those of
S. Drummondii, and the hybrid seems to resemble S. flava,
or to be even stronger.

The wing is narrow in these three forms. In 5. flava it con-
tinues undiminished up to the rim, while in S. Drummondii
it narrows down to a mere ridge, some distance below the rim.
5. Moorei has the wing narrowing down to a mere ridge, but
closer to the rim than in 5. Drummondii.

The lids in these types form an interesting series. That of
S. flava is ovate-cordate, with a strong median process, and a
straight margin. 5. Drummondii has an orbicular lid with a
blunt apex, and a wavy margin. In 5. Moorei the lid is nearly
intermediate between ovate and orbicular, with a slightly wavy
margin, and a tip process not so pronounced as in 5. flava.

In color relation the hybrid is remarkably intermediate.
The typical 5. flava is pale green, excepting for the red throat
marking. (Variations: purplish-veined lid, or with an entirely
purple pitcher.) The pitchers in S. Drummondii are green
below with white areolations over the lid and upper tube; the
areas between the areolations are marked with red veinings.
Ruddy forms of this occur. Not uncommonly there is a pale
form with green pitcher below, while, above, the areolations

20 Russell— Comparison of the Structure of Hybrid

are so enlarged that they become confluent, and the whole
upper portion appears dazzling white. To this variety has
been given the name alba. The typical hybrid shows a green
pitcher with white areolations over the lid and upper tube;
together with the red throat of S. flava, which is here lessened
in intensity. There is also a hybrid (n) noted with the white
variety of S. Drummondii named above. In it the hybrid is
pale, with pronounced areolations, as in S. Drummondii. Where
the ruddy varieties are crossed the hybrid is very beautifully
colored. The throat and rim are velvety red, with white areo-
lations and crimson veining over the lid and pitcher.

Outer Surface of the Lid

The outer epidermal cells of S. Drummondii (PI. Ill, fig. n)
are slightly angular-rounded in shape. The cell walls project
upward into a slight rounded papilla. This has been noted
by Macfarlane (28). The stomata are grouped in special tracts
above the veins between the areolations on the lid. The glands
are more numerous over this special region also, but are sparsely
scattered over the entire surface. The hairs are short and
blunt above the veins. Around the window areas the hairs
are longer.

In 5. flava the cells are longer than broad, with wavy walls.
The surface of the epidermis is flat, the hairs, stomata, and
glands are quite evenly distributed over the entire surface of
the lid. There is no suggestion of grouping of the stomata in
any special region.

The epidermal cells in the hybrid (PI. Ill, fig. 12) are squar-
ish, with slightly wavy walls. There is very little or no swelling
of the upper cell surface. The stomata are distributed in tracts,
but these areas are wider, and the stomata are less closely
grouped. There are spaces devoid of stomata as in S. Drum-
mondi, where the glands are sparsely distributed. The hairs
are long and strong over these areolations.

Inner Lid Surface

On the inner surface of the lid of 5. Drummondii (PI. IV,
fig. 18) the cells are wavy walled, measuring .06 x .05 mm. to
.05 x .04 mm. See PI. IV, fig. 16 for the appearance of the
lid cells of S. flava, already described above in detail. The cells
of the lid of the hybrid are intermediate in shape (PI. IV, fig. 19).

Sarracenias with that of Their Parents 21

The stoma ta, as in the case of the previous hybrid, form
an interesting series. S.flava has 7.25 per field, while 5. Drum-
mondii has very few — only .59 per field. The hybrid is inter-
mediate, with 3.4 per field. The size of the stomata varies
in these forms. In S. flava they are .035 by .03 mm., in S.
Driimmondii they are large, .045 X .04 mm. In the hybrid
the size is intermediate — .04 X .035 mm.

The glands present the same curious phenomenon as before;
that is, there are fewer in the hybrid than in either parent.
5. flava has 3.4 per field, S. Drummondii 4.2, while S. Moorei
has but 2.4.

The hairs of S.flava vary from .1 mm. to .3 mm., with .2 mm.
as the average length. 5. Drummondii has remarkably long,
fine hairs. There are present besides small fine hairs. The
length varies from 1 .98 mm. in length to .44 mm. The average
length is about 1.07 mm. with 1 per field. As in the former
hybrid considered, .S. Catesbaei, the S.flava parent exerts strong
influence on the size of the hairs. The hairs of S. Moorei vary
greatly in length; very few are over 1 mm. long, however. The
average length is about .5 mm.; which is close to intermediate
between the species. They are more frequent than in S. Drum-
mondii, 2.3 per field, which is many times less than the hairs
of S. flava. It seems that there is an inability in the hybrid to
reproduce to any extent the numerous hairs of the 5. flava type.

Conducting Surface

The conducting surface in these types reaches to the detentive
surface without interruption. In S. flava (PI. V, fig. 23) the
cells were .07 mm. long, with a tip process .03 mm. in length.
In S. Drummondii (PI. V, fig. 25) the tip is longer, .05 mm. in
length, while the whole cell measures .15 mm. In the hybrid
(PI. V, fig. 26) the cell is .11 mm. long, with a tip intermediate
in size, .04 mm. The glands are grouped more abundantly
over the upper conducting region, about the rim, and imme-
diately below. Below this nectariferous area, the glands be-
come less frequent and cease entirely over the lower half of
the conducting surface. In S. Drummondii there are 3-4 per
field in this upper area; in S. Moorei, 4 or more. In S. flava,
where the glands are not so limited in distribution, there are
3 per field.

22 Russell— Comparison of the Structure of Hybrid

Longitudinal Section of the Rim

A longitudinal section through the rim region of S. Drum-
mondii shows that the rim is rolled over once as in S. flava,
but with no trace of the peculiar flattening on the under part
of the turn. The tip is in 5. Drummondii pointed and elong-
ated, with thickened cells through the whole tip. The con-
ducting cells do not appear until well around the outer curve
of the rim ; while in 5. flava they occur at the outermost edge
of the rim. In S. Moorei they appear further around the rim
than in S. Drummondii. The tip region of 6". Moorei is not
so pointed as in S. Drummondii, nor as blunt as in 5. flava; it
is between the two.

There are in 5. Drummondii two rows of thickened cells
subjacent to the conducting epidermis. S. flava has only one
row of such cells, while in the hybrid there is a complete row,
and a second incomplete row.

The mesophyl of all three types is compact, excepting for
the regions of the areolations in S. Drummondii and S. Moorei.
Here the tissue is loose, with many air spaces. In S. Drum-
mondii the cells of these areas are devoid of contents, but the
tissue of this region is as deep as that of the non-etiolated por-
tions. In 5. Moorei, in young leaves, these areas are not en-
tirely devoid of chlorophyl, so that they appear light green
rather than pure white as in S. Drummondii.

Below the mesophyl there is in S. Drummondii a reversed
palisade of 3-4 layers of cells ; in S. flava there are 3-4 layers ;
in 5. Moorei, 3 complete and regular layers.

The outer epidermis of S. Drummondii is swollen into peculiar
pointed papillae. These become especially pronounced over
the areolations. S. flava has a smooth epidermis, but in S.
Moorei a trace of such swellings as are present in S. Drummondii
occur. They are here more rounded than in S. Drummondii.

Detentive Surface

On the outer part of the conducting and upper detentive
areas of 5. Drummondii there are stomata present. They are
very sparsely scattered and rather difficult to find. About
them is a group of special cells. In counting a hundred fields
for detentive hairs, four such stomata were found. In S. Moorei,

Sarracenias with that of Their Parents 23

two stomata were found in a hundred fields. The number here
is perhaps not to be taken as an average; but one may con-
fidently assert that there are fewer stomata present in the hybrid
than in S. Drummondii, although they may or may not be
intermediate in number.

The hairs of the detentive surface are about as numerous
in 5. Drummondii as in S. flava, 5-6 per field. In S. Moorei
the hairs are more numerous — 7-8 per field. In length, as
before, the hybrid is intermediate: 5. flava .37 mm. long, S.
Drummondii .54 mm., 5. Moorei .45 mm. The epidermal
cells from which they spring are straight walled cells, char-
acteristic of detentive surfaces.

Outer Epidermis of the Pitchers

The epidermal cells of the outer surface of the pitcher tube
are swollen as on the upper lid portion; excepting over the
region of the veins, where straight walled, much thickened cells
take their place. Stomata are very numerous, 10 or more to a
field; glands are infrequent, one to a field; and hairs are absent.
In S. Moorei the epidermal cells are slightly swollen; the stomata
are more numerous than in 5. Drummondii, 12 per field; glands
are 1-2 per field; no hairs are present, as in 5. Drummondii.

T. S. Tube at Base of Pitcher

The cells of the outer epidermis of 5. Drummondii are so
swollen that the stomata appear sunken, instead of raised
above the epidermal level as in S. flava. There are here but
three layers of false palisade. The mesophyl is deeper and
looser than in 5. flava with peculiar bands of cells rich in con-
tents forming a complete ring about the pitcher tube. There
are two layers of sub-detentive cells, with tannin present.

The tube sections of 5. Moorei resemble those of S. Drum-
mondii closely because of the presence here also of the peculiar
bands of parenchyma cells in the mesophyl region. There are
two layers of reversed palisade, and two rows of subdetentive
cells with peculiar aggregated contents, noted above in 61. flava.

The "ridges" above the veins are more pronounced in 5.
Drummondii than in S. flava. S. Moorei shows an intermediate
amount of thickening in this region.

24 Russell— Comparison of the Structure of Hybrid

Set 3. S. Sledgei, S. Drummondii, and S. areolata

S. Sledgei (PI. II, fig. 6) has an upright, slender pitcher,
slightly inflated in its upper half, slightly constricted at the
rim. Average pitchers are 55-65 cm. high. The pitcher of
S. Drummondii (PI. I, fig. 4), it will be remembered, increased
gradually in width, up to the rim. S. areolata (PI. II, fig. 7)
is about as high as 5. Sledgei— 55-65 cm. The rim is not con-
stricted as in S. Sledgei, nor is it so expanded as in 5. Drum-
mondii. The fused lamina extends but little in these three
forms. It is about 4 mm. wide in 5. Sledgei, while in 5. Drum-
mondii it is very narrow — only 2 mm. wide. S. areolata shows
it about as in S. Drummondii, or slightly wider.

The lid in 5. Sledgei is ovate-cordate, with a straight margin.
In its median portion it is prolonged into a tip. S. areolata
shows a lid intermediate in shape between the ovate lid of
5. Sledgei, and the orbicular frilled lid of 5. Drummondii. There
is a slight tip process.

The pitchers of S. Sledgei are green, veined with crimson
over the upper portion and the lid region. There is also a
ruddy variety with deep crimson lid and upper tube. S. Drum-
mondii, of the type described as one of the parents of S. Moorei,
has a green pitcher marked on the lid and upper tube with
white areolations and reddish veinings. S. areolata has more
pronounced reddish purple vein markings than S. Drummondii.
The white areolations are reproduced over the upper lid and
pitcher areas, but are fainter than in 5. Drummondii. Where
the 5. Drummondii parent is of the atropurpurea type, the
hybrid is very richly colored with purple lid and white mark-
ings.

Outer Lid Surface

In S. Sledgei (PI. Ill, fig. 13) the epidermal cells are longer
than broad, with angular walls. There are numerous stomata
distributed equally over the epidermal surface, averaging six
to a field. Glands are sparse — one to a field on this outer sur-
face. There are but few hairs, and these are extremely short.
In S. Drummondii (PI. Ill, fig. 11) the cells are rounded, swollen;
glands and stomata are distributed in limited areas between
the areolations. In the hybrid (PI. Ill, fig. 14) the epidermal
cells are inclined to have rounded walls, slightly swollen into

Sarracenias with that of Their Parents 25

papillae. The stomata and glands are grouped in special tracts
as in 5. Drummondii, but they are wider and less well defined.
There are strong hairs scattered abundantly over the lid sur-
face, especially around the margins of the window areas. They
are much stronger than in either parent.

Inner Lid Surface

The epidermal cells of the inner lid surface of 5. Sledgei (PI. IV,
fig. 20) are longer than broad, and wavy-walled. In 5. Drum-
mondii (PI. IV, fig. 18) they are about as long as broad, with
walls slightly wavy. S. areolata (PI. IV, fig. 21) has cells re-
markably intermediate in size and shape.

Stomata are sparsely distributed over the inner lid region
of 5. Drummondii, where the average per field is .59. In 5.
Sledgei they are quite abundant — 3-4 per field. In the hybrid
they are quite intermediate in number — 2-3 per field.

The hairs of the lid in S. Drummondii are variable in length.
They exceed in their greatest length the hairs of S. purpurea,
reaching almost 2 mm. in length, while the shorter ones are
only .3 mm. long.

Itf 5. Sledgei on the other hand the hairs are of a uniform
length. They are stout and short, varying slightly from .55
to .88 mm. long. None are as short as the shorter of those of
S. Drummondii, .3 mm. In S. areolata the hairs vary greatly
in length after the fashion of the 5. Drummondii parent. There
are hairs present as short as the shortest of S. Drummondii,
but the longest hairs are a little over half as long as the longest
hairs of 5. Drummondii — 1 mm. Hairs .7-9 mm. in length
are the most numerous. That is the influence of the S. Drum-
mondii parent is shown in the tendency to wide variation in
length, and in the lengthening of all of the hairs. The glands
in the three forms are fairly numerous — 4 in S. Drummondii,
4 in S. Sledgei, and 4 in S. areolata.

Conducting Surface
Conducting surface cells here do not show the exact relation
shown in the other hybrid. In S. Drummondii (PI. V, fig. 25)
the cells were large and measured .15 mm. in length. In S.
Sledgei (PI. V, fig. 27) the cells are about .09 mm. long, and are
narrow and oval in shape. In S. areolata (PI. V, fig. 28) the
cells are of the same length as in 5. Sledgei, but are wider and
in shape are halfway between the parent forms.

26 Russell— Comparison of the Structure of Hybrid

The glands in these three forms are most numerous about
the pitcher rim, and quickly lessen in number one inch or so
below the rim. They are scattered sparsely for a short space
below, but cease entirely beyond the upper one-third of the
conducting surface. In S. Sledgei the glands are remarkably
numerous, 7-8 per field over the upper region, near the rim.
In S. Drummondii they are more infrequent and average 3-4
per field. In the hybrid there are almost as many as in S.
Sledgei — 6-7 per field. In the extreme lower portion of the
conducting surface in all three forms are stomata scattered
very sparsely, set in a group of special cells.

Longitudinal Sections of the Rims

The rim of S. Sledgei is rolled over once in a rather loose
turn. The tip is blunt with several thickened cells. The
conducting cells appear about halfway round the outward turn.
Beneath the conducting cells is a single layer of subepidermal
cells. The mesophyl is deep and loose. There are two layers
of cells forming a false palisade below the epidermis.

In S. Drummondii there are two sub-conducting and 3-4 sub-
epidermal layers. The mesophyl is loose, especially below the
areolations.

In S. areolata the rim is intermediate. Beneath the con-
ducting cells are two layers of subepidermal cells. The meso-
phyl is deep and loose especially below the window areas. Be-
neath the outer epidermis are three subepidermal palisade
layers.

Detentive Surface

The detentive surface of S. Drummondii is, as has been de-
scribed above, composed of normal straight-walled polygonal
epidermal cells. A few normal stomata are scattered over this
surface, but no glands are present. Hairs are present, of the
normal variety, .8-.Q. mm. long, 3-4 to a field.

In 5. Sledgei this surface departs from the normal detentive
in that some of its cells are wavy walled, with very different
contents from those of the detentive epidermal cells. These
cells appear only on the upper detentive area in large groups
isolated in the midst of characteristic detentive cells. Other
cells show all transitions between straight walled detentive
surface cells, and isodiametric wavy walled cells resembling

Sarracenias with that of Their Parents 27

closely the glandular surface cells of 5. purpurea. In fact the
whole area has much the same character as the corresponding
area in the hybrid S. Catesbaei, excepting that none of these
cells appear as intrusions among the conducting surface cells,
and that they extend deep into the detentive surface. The
writer infers the probability that this region represents a primi-
tive glandular surface.

On the normal detentive surface the hairs are short, .3-4
mm. in length, but numerous, 7-8 per field. The hybrid pos-
sesses a normal detentive surface with polygonal cells and no
glands. Stomata are present. The hairs are intermediate in
size and number. They are .6-7 mm. long, with 5-6 per field.

Transverse Sections of Tube

S. Drummondii, as has been previously noted, shows in trans-
verse sections of the tube-base epidermal cells swollen into
papillae ; below these, two subepidermal layers, a loose mesophyl,
and 1-2 subdetentive layers occur.

In S. Sledgei the epidermal cells are oval. Below them are
three layers of subepidermal cells. The mesophyl is loose with
many air spaces. The mesophyl cells contain the same peculiar
contents noted in the section of 5. Drummondii. In this form
however the cells so supplied are arranged in a continuous
layer in the mesophyl tissue, below the false palisade tissue.
There are 1-2 layers of sub-detentive cells.

In S. areolata there are two layers of subepidermal cells;
and two layers of sub-detentive cells. The mesophyll is loose
as in both parents. There is a suggestion of the peculiar bands
of special storage cells, as in the case of S. Sledgei, though the
layers are not continuous.

Flowers
5. flava, S. purpurea, S. Catesbaei

The flower of 5. flava is pendulous and measures 7 cm. in
length, 13 cm. across. Those of S. purpurea are smaller, 4-5
cm. long and 9 cm. wide. In specimens at hand of 5. Catesbaei
the flowers are intermediate in size, 5-6 cm. long and 10-12
cm. wide. Other specimens seem to point to a greater robust-
ness in it than in either parent.

The bracts in all three are small, covered with honey glands
and a few stomata. The glands are distributed over the central

28 Russell— Comparison of the Structure of Hybrid

part. In color, in S. fiava they are yellowish membranous,
with green veinings. Those of 5. purpurea are reddish, or
green with a reddish margin. In S. Catesbaei the bracts are
reddish with a green tip.

The sepals of 5. flava are 4 cm. long by 3 cm. broad. They
are ovate and greenish yellow. Stomata are frequent, and the
numerous glands are massed towards the edges and tip of the
sepals on the outer side. On the inner, or morphologically
upper, surface the glands and stomata are less frequent than
on the outer side, but distributed in the same manner.

In S. purpurea the sepals are 3 cm. long, ovate as in 5. flava
and red in color. Glands are not so numerous as in 5. flava,
but distributed along the margins as above. The sepals of
•S. Catesbaei are 4 cm. long, ovate, green tinged with rosy pink,
or with red veinings. Glands and stomata are more numerous
than in S. purpurea, distributed as before.

The petals in 5. flava are 7-8 cm. long, with the proximal
portion not wider than the distal expanded part. The expanded
portion at its lower third forms a cuneate tip characteristic of
5. flava. In color the whole petal is a flavous yellow, the pig-
ment being due to yellow chromoplasts. The constricted por-

Fig. 29. Outline drawings of petals of Sarracenia flowers X }i. See text

for description.

tion of the petal (fig. 29, b) is rolled over rather deeply. Glands
and infrequent stomata are distributed alightly behind the
tip of the petal, and along several of the median veins. At
the region of the constriction they become very numerous.
In this area too, the epidermal cells become swollen into rounded

Sarracenias with that of Their Parents 29

papillae. Above the constricted area there are no stomata,
nor glands, nor papillate swellings. On the inner surface, the
distribution of stomata and glands is practically as above,
excepting for the presence of several rather long, fine hairs
along the median veins of the constricted area. S. purpurea
has petals 4-5 cm. long. The proximal portion of the petal is
not wider than the distal section. The constriction between
them is very shallow (fig. 29, a), and the sides of the petals
scarcely rolled back. The distal portion of the petal is ovate.
In color, the petal is crimson without, much lighter within.
The base of the petal is pale greenish or whitish. Glands are
numerous over the central part of the petal, back as far as the
constriction. In 5. Catesbaei the petals are intermediate in
length, 5-6 cm. long. The upper portion is slightly more
rounded than in S. purpurea; the constriction is more pro-
nounced than in the latter (fig. 29, c). The color is inter-
mediate. Above, the petal is white. The lower portion of
the petal is rose-pink. That is, the purple-pink dissolved
pigment in 5. purpurea is diluted and weakened by the chromo-
plasts from 5. flava. Glands are present at the tip, and along
the median veins only to the constricted areas.

The stamens in all are variable in number — from 50-60.
They are formed by the breaking up of each of the 10 staminal
primordia into 5-8 lesser primordia.

Pistil. The umbrelloid style is provided with numerous long
hairs, pointing generally toward the 5 stigmatic knobs. The
stigmatic hairs are short and stout, usually bent with the tip
often swollen.

In S. flava, the stigmatic hairs are short and but slightly
curved, with swollen tip.

In S. purpurea, the hairs are longer, and more decidedly
curved. The hairs of S. Catesbaei are longer than those of
>S. flava, yet not so long as those of 5. purpurea. The stylar
hairs are abundant and long in S. flava. In S. purpurea the
hairs are of about the same length, sparse within, absent with-
out. In S. Catesbaei, however, the hairs are very short.

S. flava, S. Drummondii, S. Moorei

In S. Drummondii, the flowers are pendulous, as in S. flava,
are of about the same size, 6-7 cm. long, and 14 cm. across their
greatest width. The bracts are red. S. Moorei has usually,

30 Russell— Comparison of the Structure of Hybrid

at least in all the greenhouse specimens, a larger flower than
those of the parents. They are 7 cm. long and 13-15 cm. across.
The bracts are reddish green.

The sepals are deep crimson in S. Drummondii, ovate in
shape, measuring 4-5 cm. in length. Those of 5. Moorei are
of like size, but show an intermediate coloring and shape. They
are reddish pink above, and yellow beneath, and are rather
narrower than the sepals of S. flava (35). Stomata and glands
appear as before at the tip and along the sides of the sepals.

The petals are interesting in relation in this series. They
are 6-7 cm. long in 5. Drummondii (fig. 29, d) and have a basal
segment which is rhomboidal in shape, and much wider than
the "banner" below. The margin of the petal is rolled back
strongly at the constriction ; the banner portion is oval in shape
and the whole petal is colored a rich deep crimson, due to a
dissolved crimson pigment, except at the extreme base, where
the petal is green. The epidermis of the petal, like the outer
epidermis of the pitcher, is swollen into papillae, which become
especially prominent in the constricted area. There are few
glands scattered at the lower part of the petal, and up the me-
dian veins. There are none in the constricted area of the petal,
where they are so numerous in 5. flava.

In 5. Moorei (fig. 29, e) the petals are intermediate in color
and shape. The basal portion is wider than in S. flava and nar-
rower than in 5. Drummondii. The lower portion of the petal
is pointed, somewhat suggesting the cuneate tip of 5. flava, but
rounds out above more like 5. Drummondii. In color the
petals are light at the base — a yellowish green, becoming a
peculiar maroon color below the base on the outer side of the
petal and yellow within, due to combined crimson dissolved
pigment and yellow chromoplasts of S. Drummondii with the
yellow chromoplasts of S. flava. The glands are less numerous
than in S. flava, but are distributed in the same way, toward the
tip, up the median veins and over the constricted area.

In S. Drummondii, the style is flushed with crimson over
its outer surface; while in S. flava it is yellow. In S. Moorei
the style is yellowish, with faint red markings.

The hairs of the umbrelloid style of S. Moorei are longer and
stronger than those of the parents.

Sarracenias with that of Their Parents 31

S. Drummondii, S. Sledgei, and S. areolata

The flowers of 5. Sledgei are nodding, but the petals are
spreading in this form, instead of drooping and hanging out as
banners, as in the other types so far considered. They are
rather large and showy, 5-7 cm. long, 11-14 cm. wide. Those
of S. areolata exhibit somewhat the same spreading habit as in
S. Sledgei. The flowers are 5.5-6 cm. long, 13 cm. across.

The bracts in S. Sledgei are pink with a greenish tip; those
of 5. areolata more red, due to the S. Drummondii parent.

The sepals of 5. Sledgei are greenish, fading to yellow. They
are ovate with blunt tips and measure 3-4 cm. long. Those
of the hybrid are longer, 4-5 cm., and are cordate in shape.
They are greenish with a red margin. Glands are present on
the tip and along the margins of the sepals.

The petals of 5. Sledgei (fig. 29, f) are about 6 cm. long, with
a wide basal segment, a deep constriction, and a much rounded
"banner." They are pale lemon-yellow in color, fading to
white. Glands are very sparsely distributed, one or two at
the tip and several up the central vein. They do not appear
above the constriction. As before, the cells of this portion of
the petal, on the outer surface, are swollen into slight rounded
papillae.

In 5. areolata (fig. 29, g) the petals resemble 5. Sledgei in
that the shape of the banner is decidedly rounded. The basal
portion is wider than in S. Sledgei, approaching to rhomboidal
as in S. Drummondii. In color they are intermediate between
the crimson of S. Drummondii and the yellow of S. Sledgei.
The color is rather more rose-pink than red on the outer surface,
while within the petals are pale yellow. There are no glands
present on the petals, at the constriction. The epidermal cells
are prominently papillate, though not so strongly as in S. Drum-
mondii.

The umbrelloid style is pale yellow in S. Sledgei, while in 5.
areolata it is more greenish in color, with red veining.

The hairs of the inner side of the style are short, while in
S. areolata they are remarkably long and strong, exceeding even
those of 5. Drummondii. The hairs of the stigma of S. Sledgei
are short and curved, while those of the hybrid are longer and
much less curved.

32 Russell— Comparison of the Structure of Hybrid

Glands

The structure of the glands on the pitcher surfaces, spoken
of in this paper, is in general identical. Brief descriptions
have been given, and several figures by Goebel (32), Fenner
(31), and others. There are above, appearing on the epidermal
surface, two central cells, heavily thickened, surrounded by
from 4 to 6 companion cells. The cap cells, or the central
cells, are wedge-shaped, and run down between the surrounding
cells into the second or third tier of cells constituting the gland.
The whole structure is then flask-shaped, and is composed of
16 or more cells, extending down into the mesophyl, but sep-
arated from that tissue by a cuticular lamella, with reticulate
thickenings. The cap cells mentioned above stain differently
from the surrounding cells, taking gentian-violet stain deeply,
while the surrounding cells absorb safranin. Further, the
surrounding cells give a marked positive reaction for tannin,
with ferric iron chloride. In view of the above, it seems very
probable that the cap cells are the secretory cells, and that the
tannin present is the source for the sugar in the secreted liquid.

Though the structure of these glands has been considered
briefly, there has been but one paper tracing their develop-
ment— that of Fenner (31). He describes their origin from
an epidermal initial cell which divides longitudinally and trans-
versely to form a two-tiered group of cells. But next he con-
siders that a split occurs between the upper cells, and that this
split deepens until it involves the lower layer also. Into this
"pit," he says, the secreted honey is poured. The writer finds
no evidence of any such splitting. In fact the pit he mentions
occupies the position of the cap cells. What occurs is this:
The gland initial divides longitudinally and transversely as
indicated, forming a two-tiered structure of four cells, two
above, two below. From the two upper cells are cut off two
smaller cells after the fashion of guard cells. These two cells
are the cap cells. The two upper cells redivide to form the
4-6 surrounding cells. The second tier below divides trans-
versely and longitudinally into 8-16 cells to form the second
and third tiers. While this growth is going on, the two cap
cells are pushed to a central position, and are flattened against
each other until they assume the characteristic wedge shape (32).

Sarracenias with that of Their Parents 33

Ovarian Gland Structure

The above description applies to the glands present on all
the vegetative parts, as well as on the bract, sepals, the petals
and umbrelloid style. But over the tubercles on the surface of
the ovary are still more complex glands. Macfarlane (28)
gives the following description of the nectar secretion: 'The
epidermal cells of the ovarian surface have undergone repeated
divisions, and have swollen out into minute glassy beads or
tubercles from which a quantity of rich nectar exudes before,
during, and for some time after, blossoming. This, as we will
show, is evidently of great use in the pollination of the flower."
"When a flower has nearly opened the stamens begin to dehisce
and as the blossom has a pendulous position the pollen from the
stamens is showered down into the umbrelloid style-cavity below.
But about this time the warted bead-like ovarian surface exudes
large drops of sweet juice, which increases in quantity as the
stamens continue to dehisce, till it oozes down among the fila-
ments and anthers, washing with it the pollen-grains. It then
accumulates in the umbrelloid cavity, and forms there a nectar-
bath of pollen."

When serial sections of the ovarian wall of an opened flower
are studied, the source of the abundant honey is seen to be in
the large nectar glands situated at the bases, and along the
sides, of the tubercles mentioned above by Macfarlane. The
glands are more numerous over the tubercles of the lower half
of the ovary. The tubercles over the upper surface of the ovary
are devoid of glands. Following is an account of the mode of
origin of the tubercles and glands:

In a very young flower, where the anthers show the pollen
mother cell stage, the ovary is as yet very small, and its epi-
dermis smooth. The ovarian tissue is, however, growing rapidly
at this period, and division figures are frequent. They indicate
a longitudinal division for the epidermal cells invariably, while
in the subjacent tissue both longitudinal and transverse
divisions are common. At this time both petals and sepals
possess fully formed glands and stomata. The ovarian wall as
yet shows no trace of either. The epidermis continues to
divide more frequently than the tissue below. This excessive
division is somewhat localized along the ovarian wall, so that
there results here and there a slight swelling. The layer im-

34 Russell — Comparison of the Structure of Hybrid

mediately below the epidermis becomes differentiated also from
the ovarian tissue. The cells become enlarged and filled with a
peculiar content present in the epidermis also. This layer
shows rapid division and pushes up into the swelling. It is
impossible at this time to trace the gland initial cell. At this
time, the pollen mother cells are in the anaphase of the first
or heterotypic division.

As the epidermal cells divide, the swellings become more
pronounced. Two subepidermal layers become involved. The
outer cells of the tubercles become enlarged and oblong while
the cells between the swellings become compressed and small.
The first gland initials are found at this stage. They are usually
situated at the base of the swellings and consist of one or two
cells with a large nucleus and richly granular protoplasm. At
this time the pollen grains are in the tetrad stage, and the mega-
spore mother cell is distinguishable.

The swellings become deepened as the flower matures and
press together tightly so that they assume a squarish or oblong
shape (fig. 30). At this time the glands are fully developed.

Fig. 30. Longitudinal section through mature ovarian wall, X 100. G =
gland in section; E = epidermis with waxy layer on outer surface.

They have the same fundamental structure as the simpler
glands on the leaves, etc.; that is, they possess a set of 2 or
frequently 3-4 central cells running down into a mass of small
cells below.

The gland is irregularly spherical in shape, and is composed
of many more cells than the glands on the pitchers. There
may be 60-80 cells concerned in the formation of the ovarian

Sarracenias with that of Their Parents 35

glands. They lie embedded in the tissue of the papillae, for
the most part at their base. The honey secretions are poured
out into the tiny crevasses between the papillae, and are drawn
upwards by capillarity to the lower part of the ovary, there
to collect about the base of the filaments as has been noted by
Macfarlane above.

The papillae, mentioned as being nectariferous (28) in them-
selves, are oblong masses of fundamental tissue. The outer-
most or epidermal layer and a subepidermal layer become en-
larged, and regular, forming a noticeable bounding layer. These
two outer layers contain a peculiar substance from the earliest
stages of their differentiation. They have on their outer sur-
face a coating of wax which gives with alkanna and Soudan 1 1 1
a characteristic reaction. The tissue of the papillae is richly
supplied with oil globules. The contents of the cells of the
bounding layers stain deep brown with iodine and potassium
iodide, indicating perhaps that an alkaloid is present. Its
position and abundance may indicate the presence of such a
protective substance. An alkaloid has been mentioned as being
present in the family by Porcher (34). Sarracenin was the
name temporarily applied to it, and it was thought to have
medicinal value for stomach troubles and smallpox. Such an
alkaloid, however, is not mentioned in such a treatise as that of
Winterstein and Grier. Several diagnostic tests for alkaloids
were applied besides iodine in potassium iodide. Nitric acid
on sections dampened with potassium hydroxide in alcohol
gave a rather deep orange color, with here and there a decided
pink tint in the bounding layers. No reaction occurred with
platinic chloride. On adding ammonia, a decided bright green
color was obtained.

In view of the fact that these papillae have their outer walls
heavily thickened and covered with a protective waxy coating;
and are supplied with abundant nectariferous glands over the
lower portion, it is evident that the nectar is secreted by the
glands and not by the papillae as a whole.

Conclusions

It will be seen from the above description that the hybrid
forms, in comparison with their parents, are intermediate in
relation in almost all details.

36 Russell— Comparison of the Structure of Hybrid

In size it has been shown that the hybrids are generally inter-
mediate, though S. Moorei and S. areolata frequently incline
to show a characteristic increase in vigor over the parent types.

In shape the intermediateness is particularly well shown in
the first hybrid, 5. Catesbaei (fig. 3), where the diverse shapes
and habits of the two parents are so neatly blended. The
parents of the other forms do not present so diverse an appear-
ance; but in small details, such as the shape of the mouth open-
ing, or the width of the fused laminae, the hybrid shows an
intermediate character.

In all the blending of the parental lid shapes is shown. Where
a form with a frilled margin — as in 5. Drummondii or S. pur-
purea— is crossed with a form with a straight margin, the result-
ing hybrid has a lid with a frilled margin, but more loosely
wavy than in the parent. When a form like S. fiava, which
has a median tip process on the lid, is crossed with a blunt
tipped form like S. purpurea, the resulting hybrid has a tip,
but much weaker than that present in 5. fiava.

In the matter of coloring, the blending is beautifully shown.
If a form with red markings be combined with a green form
showing no such markings, the hybrid will have the markings
reproduced at half the intensity of the parent.

The flowers, in the matter of their size and shape, show a
marked blending, though here too the hybrid is inclined to
be larger and more showy than either parent.

A remarkable series is shown in the comparative petal shape
and size. The figures given show how intermediate the hybrid
is in relation.

In odor also the hybrid blends the parental characters —
for instance, 5. fiava has a very decided and unpleasant odor,
while 5. Drummondii has a rather delicate sweet scent. Their
hybrid has a stronger odor than 5. Drummondii, but not at all
unpleasant.

The intermediate relation in the matter of flower coloring
in the three sets has been noted. Quite noticeable is the pecu-
liar maroon color obtained in the hybrids with 5. Drummondii
as a parent.

All of these examples of blending parental characters seen
in the hybrids are such as any one can note with the naked eye.
The intermediate relation is, however, much more intimate and

Sarracenias with that of Their Parents 37

exact. It extends to microscopic details in structure to be
noted below.

The epidermal cells of the outer surface of the pitcher and
of the inner lid surface show a blended appearance in the
hybrid. Take for example 5. Moorei (fig. 12), whose cells are
intermediate between the rounded epidermal cells of S. Drum-
mondii (fig. 11) with their pronounced papulation; and the
wavy walled cells, without papillae, of S. flava (fig. 9).

The epidermal cells of the conducting surface of the hybrids
are remarkably intermediate between the parents. Here there
are measurable differences in length of cell and tip of process,
and it has been shown that the hybrid lies exactly between the
parents in length relation, breadth, and tip length.

On the outer and inner surfaces where stomata are present,
the hybrid shows the number present in a given field to be an
exact arithmetical mean between the numbers present in the
parents. In regard to their distribution it has been noted that
in other species the stomata are equally distributed over the
surface, while in S. Drummondii they are limited to special
tracts between the window areas. The hybrids with S. Drum-
mondii show the stomata distributed in wide tracts (figs. 12, 14).

The unicellular hairs distributed over the various surfaces
show several interesting relations.

On the detentive surface, the hybrid presents an arithmetical
mean between the parent types in number and length of hair in
each set. Over the outer surface of the pitcher, where the
hairs are irregularly scattered, the relation is somewhat ob-
scured. But in S. flava (fig. 9), S. Catesbaei (fig. 10) , 5. pur-
purea (fig. 8) one may see how the hybrid shows a variety of
hair lengths inherited, and an intermediate degree of blending
in the hairs.

On the inner lid surface, the hairs of all except two species
show great variability in length. The hybrids with either of
these two forms S. flava and S. Sledgei as a parent, having fairly
uniform hairs, with any of the other forms, having variable
lengths, show hairs of all lengths reproduced, but fewer in
number and with the longest hairs much reduced.

This behavior of hair-length inheritance has been noted be-
fore in other plants (29) so that this seems to be a rule of inheri-
tance for hair length and number.

38 Russell— Comparison of the Structure of Hybrid

In relation to the actual structure of the pitcher, longitudinal
sections of the rim have been compared. It has been noted
that, in the rolling, in the shape of the tip, in the amount of
thickening at the tip, the hybrid is intermediate in character.

In all transverse and longitudinal sections of the pitcher,
the number of layers of false palisade, of subepidermal cells,
and the depth and character of the mesophyll, have all indi-
cated an intermediate relation in the hybrid.

In the amount of thickening, either in sclerenchymatous
tissue about bundles, or in the number of striations on the
hairs of the inner lid surface, the hybrid is invariably inter-
mediate.

Bisexual Hybridity

Some apparent variations from the exact blending should be
noted here. On pages 16 and 17 there is an account of the
inability of the glandular surface of S. purpurea to blend with
the conducting surface of S. flava. The resulting mosaic effect
is described. This obtains also in the hybrid of S. purpurea
with S. Drummondii, and has been noted in the hybrid with
5. minor. The inability is probably due to the fact that S.
purpurea has evolved far in advance of all other forms, except
possibly 5. Sledgei. It would be interesting to know if of all
the forms this could perfectly blend with 5. purpurea when
crossed. Unfortunately such a hybrid has not been found.

In the hybrid of 5. Sledgei with S. Drummondii, the develop-
ing glandular area is completely lacking; the impulse is too
weak for transmission.

Other peculiarities of relation in parent and hybrid types
might be illustrated by the relation shown in comparisons of
counts obtained for glands. In every case practically the
hybrid shows fewer glands than either parent. It is almost
impossible to obtain a hundred or more counts from a given
surface, and compare it with a hundred similar counts from
exactly corresponding surfaces. The glands are grouped in
S. flava for instance around the rim, and down the median
back portion of the pitcher. They extend halfway down the
length. In S. Drummondii they are massed about the rim
too, but are very sparse just below. If the total number of
glands could be counted and compared, they would undoubt-

Sarracenias with that of Their Parents 39

edly show a perfectly exact relation. This is upheld by the
fact that where a surface without glands (in 5. flava) is crossed
with a surface where the gland number is uniform (glandular
of S. purpurea) the hybrid shows an exactly intermediate num-
ber. In the case too where they are massed in the same regions
in parents and hybrids, as in S. Sledgei, S. Drummondii, and
S. areolata, the hybrid shows a blended relation.

The comparison of these three hybrid plants with their parents
in all details of structure gives an overwhelming mass of evi-
dence for exact blending, which extends to the most minute
details. It surely points to some exact relation in molecular
structure of the hybrid plant, extending even to the amount
of thickening laid down in a cell wall, the size of the starch
grains, or the size of a chloroplast.

The writer wishes to acknowledge the kindness and coopera-
tion of Dr. J. M. Macfarlane in the preparation of this paper.
Thanks also are due to Mr. W. R. Taylor, who made many
of the microphotographs; and, through the kindness of Mr.
Frank M. Jones, the writer was able to examine fresh flowers,
which he sent from Alabama, Mississippi, and Florida.

BIBLIOGRAPHY

1. Tournefort, J. T. Institutione Rei Herbariae, 3rd Ed. (1719)-

2. Linnaeus, C. Species Plantarum, 1st Ed. (i753)-

3. Clusius, C. Rariorum Plantarum Historia (1601).

4. Parkinson, John. Theatrum Botanicum (1640).

5. Josselyn, New England Rarities Discovered (1672).

6. Plukenet, L. Phytographia, Part III, Tab. 152 (1692).

7. Plukenet, L. Amaltheum, Tab. 376 (1705).

8. Ray, John. Historia Plantarum, Vol. II (1688).

9. Walter, Thos. Flora Caroliniana (1788).

10. Bartram, Wm. Travels (1791).

11. Michaux, Andre Flora Boreali America (1803).

12. Croom, H. B. Catalogue of Plants (1737)-

13. Anonymous. Flore des Serres (1851).

14. Chapman, A. W. Flora of the Southern United States (i860).

15. Hooker, W. J. Gardeners' Chronicle, Vol. IX, p. 260 (1874).

16. Boulger, Prof. Gardeners' Chronicle, Vol. XV, p. 627 (1881).

17. Macfarlane, J. M. Roy. Hort. Soc. Conf. Genetics, p. 155 (1906).

18. Masters, M. T. Gardeners' Chronicle, Vol. XV, p. 816. Vol. XVI,

p. 11, 40 (1881).

19. de Candolle, A. Prodromus, Vol. XVII (1870).

40 Russell— Comparison of the Structure of Hybrid

20. Nicholson, George.

21. Moore, David.

22. Veitch & Son.

23. Macfarlane, J. M.

24. Macfarlane, J. M.

25. Bailey, L. H.

26. Elliot, Stephen.

27. Eaton, Amos.

28. Macfarlane, J. M.

29. Macfarlane, J. M.

30. Solereder, Hans.

31. Fenner, C. A.

32. Goebel, K.

33. Harper, R. M.

34. Harper, R. M.

35. Mohr, C.

36. Shreve, F.

37. Porcher, F. P.

38. Walter, Herbert E.

Dictionary of Gardening, Vol. III.

Gardeners' Chronicle, Vol. XXII, p. 702 (N. S.)

(1874).
Gardeners' Chronicle, Vol. XXIII, p. 73» (1874)-
Pflanzenreich IV, 1 10 (1906).
Contrib. Bot. Lab., Vol. II, p. 426 (1904)-
Encyclop. Horticulture, Vol. VI (1917)-
Botany of South Carolina and Georgia (1821).
Manual of Botany (1833).
Annals of Botany, Vol. VII, p. 445 (1893).
Trans. Roy. Hort. Soc, Vol. 37 (1892).
Anatomy of the Dicotyledons, Vol. I, p. 51 (1908).
Flora, Vol. 93, p. 351 (1904)-
Pflanzenbiologischen Schilderungen, Vol. 2, p. 89

0893).
Jour. Elisha Mitchell Scientific Soc, Vol 39, No. 3

(1918).

Bull. Torrey Bot. Club, Vol. 34, p. 351 (1907).

Bull. Torrey Bot. Club, Vol. 24, p. 23 (1897).

Bot. Gazette, Vol. 42, p. 107 (1906).

Resources of the Southern Fields and Forests

(1869).
Genetics, An Introduction to the Study of Heredity,

p. 12 (1913)-

EXPLANATION OF PLATES

Plate I.

Fig. 1.

5.

purpurea.

Fig. 2.

5. flava.

Fig. 3-

S.

Catesbaei.

Fig. 4.

S.

Drummondii

Plate II

Fig. 5-

S.

Moorei.

Fig. 6.

S.

Sledgei.

Fig. 7-

S.

areolata.

(The above photographs were made by W. R. Taylor from specimens in the
Sarracenia House at the Univ. of Penna. Bot. Gardens.)

Plate III. Micro-photographs of outer epidermis of lids of pitchers X 70.
Fig. 8. 5. purpurea.
Fig. 9. S. flava.
Fig. 10. 5. Catesbaei.
Fig. 11. 5. Drummondii.

Sarracenias with that of Their Parents 41

Fig. 12. S. Moorei.
Fig 13. 5. Sledgei.
Fig. 14. 5. areolata.

Plate IV. Micro-photographs of the inner epidermis of lids of pitchers X 50.

Fig. 15. S. purpurea.

Fig. 16. S. flava.

Fig. 17. S. Catesbaei.

Fig. 18. 5. Drummondii.

Fig. 19. S. Moorei.

Fig. 20. 5. Sledgei.

Fig. 21. S. areolata.

(Photographs by W. R. Taylor.)

Plate V. Micro-photographs of epidermal cells of the conducting surface
X 500.

Fig. 22. S. purpurea.

Fig. 23. 5. flava.

Fig. 24. 5. Catesbaei.

Fig. 25. S. Drummondii.

Fig. 26. S. Moorei.

Fig. 27. 5. Sledgei.

Fig. 28. 5. areolata.

(Photographs by W. R. Taylor.)

A Comparative Study of the Structure and

Saprophytism of the Pyrolaceae and

Monotropaceae with Reference

to their Derivation from

the Ericaceae

Margaret W. Henderson, B.S., M.A.

[Thesis presented to the Faculty of the Graduate School in partial fulfillment of the
Requirements for the Degree of Doctor of Philosophy.]

CONTENTS

PAGE

Introduction 43

Historical 48

Methods and Materials Used 5°

The Underground Root and Stem Systems 51

The Root 53

The Rhizome 64

The Ascending Axis 64

The Leaf 75

The Leaf — Microscopic Structure 78

The Inflorescence 87

The Sepals 90

The Petals 92

The Stamens 94

The Pistil 97

The Fruit and Seed 10 1

Summary 103

Conclusions 105

Bibliography 106

42

Introduction

In the families of flowering plants which show saprophytism
and parasitism there occur usually green purely autophytic
plants with typical green leaves and numerous flowers; plants
that are purely saprophytic or parasitic, with colorless scales
and a reduced number of flowers; and all gradations between.
In a comparative study of such typical families showing sapro-
phytism, i.e., the Burmanniaceae, Orchidaceae, Gentianaceae,
and Ericaceae, and those showing parasitism, the Loranthaceae,
Santalaceae, Balanophoraceae, Rafflesiaceae, Lauraceae, Con-
volvulaceae, and Scrophulariaceae, one notices a common ten-
dency in the saprophytic or parasitic members toward condensa-
tion and simplification as the saprophytism or parasitism be-
comes more pronounced. To illustrate from the parasitic group,
in the Loranthaceae the genera Nuytsia and Gaiadendron are
upright independent trees; the genus Loranthus consists of
upright shrubs with large leaves and numerous flowers as L.
Baroni Baker and L. pulcher D.C., to those with small leaves
and solitary flowers as L. microcuspis Baker and L. stocksii
Hook.; Viscum consists mainly of species of condensed habit,
small leaves often reduced to scales, and small green unattractive
flowers and more or less simplified embryos; finally Arceutho-
bium consists of reduced almost leafless parasites becoming
slightly yellowish in color with small solitary flowers. In this
family, however, the plants still contain chlorophyll. In the
Convolvulaceae the parasitism has become so great in the
genus Cuscuta that it has completely lost all traces of chloro-
phyll, except in the stems of C. reflexa, Roxb. (Hooker in Bot.
Mag. t. 6566). From this species we have gradations to others
with thick yellow stems like C. exaltata, Engelm.; others with
slender yellowish or red stems as C. epilinum Weihe and C.
epithymum Murr, pale yellow in C. arvensis Beyrich, and whitish
or pale yellow in C. cephalanthi Engelm., and finally to white in
C. epithymum var. alba. The leaves are in all cases reduced
to microscopic scales; the flowers are small but clustered to-
gether; the first or central flowers are five-parted, the lateral
ones often four-parted (15), and the embryo is so reduced that
it shows no trace of cotyledons.

43

44 Henderson — Comparative Study of Pyrolaceae and

The writer, however, is concerned with saprophytism alone.
In the Burmanniaceae, for example, the genus Burmannia shows
transitions from green leafy plants with several racemose flowers
as B. longijolia Becc. to gradually condensing forms, as B.
azurea Griff, with a rosette of tiny herbaceous membranous
leaves and one to four flowers, then to more simplified forms,
as B. tuber osa and B. Candida, and finally, most simplified of all,
to the genera Thismia and Gymnosiphon. The stems become
feeble, less green, then reddish or brownish. The green leaves
become reduced to herbaceous membranous leaves, then to
scales; the flowers become reduced in number and size; the ovary
becomes reduced from a three-celled condition with central
placenta to a more primitive one-celled state with parietal
placentas (Gymnosiphon) ; there is an increase in the number of
ovules, but a reduction in their size; the seeds are reduced in
size and structure; the reserve albumen is reduced in size and
number of cells, and the embryo from a typical monocoty-
ledonous one to a formless mass.

The writer would claim that essentially the same set of changes
can be traced in genera of the Orchidaceae and Gentianaceae.
Green autophytes, passing by gradual changes to colorless
saprophytes, occur in both of these. Now, among systematists
of the past, there has been no thought of putting the sapro-
phytic plants of the above three groups in separate families.
Why then, it may be asked, should the Pyrolaceae and Mono-
tropaceae be separated from the Ericaceae? It will be the
writer's aim in the present paper to prove, alike on morpho-
logical and physiological, as well as taxonomic grounds, that
these three families all show so close a relationship that to view
them as separate families is unnatural and artificial.

Jussieu (37) considers all ericaceous plants under the two
orders — Rhododendra and Ericae — the latter including Pyrola.
Lindley (46) places them in three orders — Ericeae, Vaccinieae,
and Pyrolaceae (including Monotropaceae). De Candolle (10)
makes four orders — Vaccinieae, Ericaceae, Pyrolaceae, and
Monotropeae. Gray (25) considers the Vaccinieae, Ericineae,
Pyroleae, and Monotropeae as suborders of the Ericaceae.
Baillon (3) considers the Pyroleae, Monotropeae, and Ptero-
sporeae as three of the eighteen tribes under the Ericaceae.
Bentham and Hooker (4) make three orders, the Vacciniaceae,

Monotropaceae with Reference to Ericaceae 45

Ericaceae (including as a tribe the Pyroleae), and the Mono-
tropeae. Drude (12) gives two families — the Ericaceae and
the Pyrolaceae (including the Monotropaceae). Britton and
Brown (6) consider as separate families the Pyrolaceae, Mono-
tropaceae, Ericaceae, and Vacciniaceae. Small (72) considers
the Pyrolaceae, Monotropaceae, and Ericaceae as distinct
families. So we see that the Pyrolaceae and Monotropaceae
are united by the majority of systematists, yet the gulf between
the two is really more difficult to bridge than that between
the Ericaceae and Pyrolaceae.

Drude in "Die Natiirlichen Pflanzenfamilien" (12) gives as
his reasons for not including the Pyrolaceae and the Mono-
tropaceae in the Ericaceae: (1) the remarkable placentation,

(2) the regular form of the seed and embryo in the Ericaceae,

(3) the lacking disc, (4) the dehiscence of the anthers, (5) the
simple pollen of the Monotropeae.

(1) In typical Ericaceae the ovary is five- or four-celled
with a central placenta. Two lobes bearing the ovules extend
into each cell of the ovary. It is generally considered to be
five-celled in the Pyrolaceae also, but the division is not com-
plete. The parietal placentae borne in on the dividing walls
fuse at the center for about half the length of the ovary. This
basal half is exactly similar to that of the Ericaceae, two pla-
cental lobes bearing ovules extending into each cell. Above
this the placentae fail to meet at the center and the upper half
becomes one-celled with bilobed parietal placentae. This con-
dition is true of C. umbellata, C. maculata, P. rotundifolia (P.
americana Sweet), P. elliptica, P. secunda* P. minor, P. chlor-
antha, P. aphylla, Moneses uniflora* of the Pyrolaceae; also of
Allotropa virgata, Pterospora andromedea, Sarcodes sanguinea,
Schweinitzia odorata, Monotropa hypopitys, M. uniflora, of the
Monotropaceae. The division Pleuricosporeae of the Mono-
tropaceae is considered one-celled in the ovary. Newberrya
is described by Torrey in the Ann. Lye. N. Y. VII 55 (1864):
"Placentae four with broad divergent lamellae which meet
adjacent edges, ovuliferous both sides giving the appearance of

* P. secunda and Moneses uniflora are almost completely five-celled.
Owing to the sunken style the distance through which the ovary is one-
celled is very short.

46 Henderson — Comparative Study of Pyrolaceae and

four exterior cells surrounding a central large one." The writer
examined material of Newberrya spicata A. Gray and N. con-
gesta (A. Gray) Torr. and found an exactly similar condition
to that in the rest of the family — i.e., that the ovary at the base
is five-celled owing to the fusion of the placentae and that
owing to a lack of fusion further up the ovary became one-celled
with parietal placentae. Pleuricospora fimbriolata A. Gray
shows the most simplified condition in regard to the ovary.
It is four-celled only at the very base for about one-sixth the
distance, then one-celled with four parietal placentae.

(2) In typical Ericaceae the seed is very small, never larger
than 1-2 mm. The seed covering is double, there is a richly
developed endosperm in which is embedded a straight embryo
which is one-third to two-thirds the length of the seed. The
embryo always shows a root, an axis, and two cotyledons. The
seeds in the Rhododendroideae-Ledeae, the tribe nearest in
character to the Pyrolaceae, are winged, very small, and con-
tain a very small embryo. These seeds are very similar to
those of the Pyrolaceae, except that in the latter the embryo
itself is very much smaller, simplified to a few cells, and with
no differentiation into root or cotyledons. The amount of
albumen in which the embryo is embedded is also reduced.

(3) There is in practically all typical Ericaceae, at the base
of the ovary, a nectar-secreting disc which may be present
as a circular ring or a crenulately lobed swelling. M tiller has
illustrated those of Arctostaphylos, Calluna, Erica, Azalea, Rho-
dodendron, and Warming those of Andromeda, Cassiope, Phyllo-
doce, in Knuth's "Handbook of Flower Pollination" (42). Drude
(12) seems to indicate that it is lacking in the Pyrolaceae and
Monotropaceae and gives this as a reason for separating them
from the Ericaceae. In C. umbellata and C. maculata it is
present as a nectar-secreting ring at the base of the ovary. In
the genus Pyrola the disc varies. According to Drude (12) it
is present as a ten-rayed nectar-secreting organ in P. (Moneses)
uniflora, but Miiller (42) states that there is no nectar secreted
by M. uniflora and figures no nectaries. The writer sees no
trace of nectaries in the material examined. Drude states
that the disc is present as ten small nectaries at the base of the
ovary in P. secunda and is absent or rudimentary in all of the
other species. The writer found very small swellings that

Monotropaceae with Reference to Ericaceae 47

appeared glandular at the base of the ovary in P. secunda, P.
chlorantha, P. aphylla, but none in P. rotundifolia, P. elliptica,
and P. minor.

In the Monotropaceae Drude says "Disc present, or replaced
by nectaries, rarely lacking." In Allotropa, according to Torrey
and Gray (77), it is minutely ten-lobed. The writer finds ten
slightly downward directed lobes; in Monotropa there are 8-10
downward directed nectaries; in Sarcodes* there are ten swell-
ings at the base of the ovary; in Pterospora Drude reports that
it is absent.

In Schweinitzia there are ten lobes between the stamens.
Drude states that in the Pleuricosporeae the disc is entirely
lacking in Newberrya, Pleuricospora, and Cheilotheca. This
appears to be true of Pleuricospora but in Newberrya the writer
finds that ten nectaries are present at the base of the ovary,
very similar to those of Monotropa. No material of Cheilo-
theca could be obtained for examination.

(4) In the Pyrolaceae and Monotropaceae the dehiscence of
the anthers follows one of these types.

1 . Apical porous — with more or less developed tubes in Chima-
phila, Pyrola, Moneses, Sarcodes, Schweinitzia.

2. Longitudinal — Allotropa, Pterospora, Pleuricospora, New-
berrya, Cheilotheca.

3. Transverse — Monotropa.

In the Ericaceae apical porous and longitudinal dehiscence
seem to be about equally distributed throughout the family
and even in the same group, i.e., in the Rhododendroideae-Ledeae
Bejaria and Ledum have apical porous, Elliottia and Clado-
thamnus longitudinal dehiscence of the anthers. Transverse
dehiscence of the anthers also occurs in the Ericaceae. The
group Arbutoideae-Andromedeae, according to Drude (12), has
pores or slits at the apex or transverse slits.

(5) The Pyrolaceae and Ericaceae have tetrad pollen grains,
the Monotropaceae simple pollen grains. However Pyrola
secunda, a typical member of the genus, has simple pollen grains
so that even this distinction does not entirely hold true.

* Drude reports no disc in Sarcodes; Oliver (58) that the disc is present.
The writer's material confirms the latter.

48 Henderson — Comparative Study of Pyrolaceae and

We see, therefore, how artificial, untrustworthy, and inter-
blending these distinctions are. There are also other great
resemblances between the Ericaceae and the Pyrolaceae and
Monotropaceae. All of the Ericaceae are shrubby rarely arbor-
escent, often sub-shrubby. The Pyrolaceae as a whole are sub-
shrubby, Chimaphila umbellate often becoming very thick,
woody; most of the genus Pyrola is sub-shrubby also, with the
exception of Pyrola or Moneses uni flora. The Monotropaceae
are essentially saprophytic herbs, but they perennate from
underground woody parts.

Historical

Owing to the complete historical references on the subject
of mycorhiza in general given by Gallaud (22) and Rayner (64),
the writer has confined herself to those dealing with mycorhiza
in the Ericaceae alone.

As early as 1842, Rylands (67) investigated the "byssoid"
substance on the roots of Monotropa hypopitys and came to the
conclusion that it was a fungus living on the roots. In 1856-
1865, Chatin (6) described M. hypopitys as a parasitic plant.
In 1873, Drude (11) investigated the roots of M. hypopitys,
coming to the conclusion that the plant is saprophytic. In
1 88 1, Kamienski (38, 39) also described and figured the con-
dition in M. hypopitys, and formulated the hypothesis that
there existed a symbiosis between the plant and the fungus.
During the last thirty-five years, it has been shown that sapro-
phytism is widespread throughout the Ericaceae proper. Frank
(19), in 1887, described the appearance of ericaceous roots
infected by fungi. He mentions particularly the much en-
larged epidermal cells, filled with knots of hyphae, the absence
of root hairs, and the reduction of the root cap. Those men-
tioned as possessing endotrophic mycorhiza are Andromeda
polifolia, Vaccinium oxycoccus, V. uliginosum, V. macrocarpum,
V. myrtillus, V. vitis-idaea, Ledum palustre, Calluna vulgaris,
Rhododendron ponticum, Azalea indica, Empetrum (included by
him under Ericaceae). He says, however, that they have not
been found in Pyrola and that in Monotropa the mycorhiza is
ectotrophic.

Monotropaceae with Reference to Ericaceae 49

In 1907, Charlotte Ternetz (77), in a paper "Ueber die As-
similation des Atmospharisches Stickstoffes durch Pilze," used
as the basis for her experiments endotrophic fungi in the roots
of Andromeda polifolia, Oxycoccus palustris, Calluna vulgaris,
Erica carnea, E. tetralix, Vaccinium myrtillus, V. vitis-idaea,
V. uliginosum. She states that, although hyphae are present
in the seed coats, no trace of the fungus can be found in any
other living part of the plant, except the roots. Rayner (63)
in an article on "Obligate Symbiosis in Calluna vulgaris" traces
the fungus from the roots through the whole plant to the seed.
She proves that the symbiosis in Calluna is obligate, for, unless
the seedlings become infected, they die. She also states that
ovarial infection is present in Ledum palustre, Rhododendron
ponticum (garden var.), Rhododendron indicum {Azalea indica,
garden var.), Leiophyllum buxifolium, Kalmia angustifolia, Pieris
fioribunda, P. japonica, Gaultheria acutifolia, Arctostaphylos
uva-ursi, Arbutus unedo, Vaccinium vitis-idaea, Pentaptergyium
serpens, Calluna vulgaris, Erica carnea. Jean Dufrenoy (14)
in "The Endotrophic Mycorhiza of Ericaceae" has reported
the presence of a fungal mycelium throughout the entire plant
of Arbutus unedo.

All these cited are, however, still green, and have abundant
leafy branches with well-formed clusters of typically ericaceous
flowers. It seems that the more simplified greens and the
yellow and white saprophytic forms have been separated from
the Ericaceae to form the Pyrolaceae and Monotropaceae, and
that these three families in the order named form a continuous
series from autophytic to completely saprophytic plants.

An aim of this paper will be to ascertain how far saprophytism
has caused gradual and traceable degradation changes similar
to those of the Burmanniaceae, Orchidaceae, and Gentianaceae.
Practically all typical ericaceous plants are shrubs, rarely trees;
some like Cassiope are sub-shrubby; in the Pyrolaceae Chima-
phila and the larger species of Pyrola are sub-shrubby; but
Moneses uniflora would hardly be regarded as other than a
herb; in the Monotropaceae all of the genera are herbaceous.
Another aim will be to find characters which unite the most
degraded saprophytes with the autotrophic Ericaceae, thereby
proving that the separation of the Pyrolaceae and Monotro-
paceae from the Ericaceae is artificial.

50 Henderson — Comparative Study of Pyrolaceae and

Methods and Materials Used

In making comparisons, care was taken to section the material
at exactly the corresponding point in each of the plants.

Sections of freshly gathered or of alcoholic material of rhi-
zomes, stems, and leaves were examined unstained in acetic
acid, or stained in safranin and methyl green and examined in
balsam.

Fresh root tips were examined in a solution of iodine in potas-
sium iodide to differentiate the fungus. Otherwise, all roots
and flowers were fixed in weak chrom-acetic acid and embedded
in paraffin. These were stained in safranin and gentian violet.

Pieces of the ascending axis of Monotropa uniflora and M.
hypopitys were bleached and macerated in a mixture of 50%
nitric acid and potassium nitrate in order to examine the epi-
dermis.

Herbarium material of the flowers was prepared for exam-
ination by the method used by R. C. McLean (52).

The following is a list of the plants used, and localities from
which they were collected, for use in this comparison:

Fresh Material

Chimaphiia umbellata — Hosensack, Pa., Analomink, Monroe
Co., Pa., Somers Point, N. J., Glendora, N. J., Woods Hole,
Mass., Greenwood Lake, N. Y.

Chimaphiia maculata — Hosensack, Pa., Crum Creek, Dela-
ware Co., Pa., Almonessen, N. J., Blackwood, N. J., Somers
Point, N. J., Woods Hole, Mass.

Pyrola rotundifolia — Hosensack, Pa., Crum Creek, Delaware
Co., Pa., Analomink, Monroe Co., Pa., Woods Hole, Mass.

Pyrola elliptica — Hosensack, Pa., Crum Creek, Delaware Co.,
Pa., Analomink and Paradise Valley, Monroe Co., Pa., Woods
Hole, Mass.

Moneses uniflora — Plants collected by Miss Mary Garley near
Claremont, N. H.

Monotropa hypopitys — Analomink, Pa., Somers Point, N. J.,
Woods Hole, Mass.

Monotropa uniflora — Analomink, Pa., Blackwood, N. J.,
Woods Hole, Mass.

Kalmia latifolia, Kalmia angustifolia, Dendrium buxifolum,
Cassandra calyculata — Clementon, N. J.

Ledum groenlandicum — Peakes Island, Maine, collected by
Miss A. M. Russell.

Monotropaceae with Reference to Ericaceae 51

Alcoholic Material

Sarcodes sanguinea — Collected by Miss Edith M. Farr in
California.

Herbarium Material

Descriptions of the plants and microscopic studies of the
flowers of Pyrola secunda, P. minor, P. chlorantha, P. aphylla,
Pterospora andromedea, Schweinitzia odorata, Pleuricospora fim-
briolata, have been made from the Herbarium of the University
of Pennsylvania. Those of Allotropa virgata, Newberrya spicata,
N. congesta, were examined from the Herbarium of the Phila-
delphia Academy of Natural Sciences.

As it was impossible for the writer to get fresh material of
the roots of Moneses nniflora, the Herbarium of the Bronx
Botanical Garden, through the kindness of Dr. F. W. Pennell,
sent several herbarium sheets for examination.

Through the kindness of Mr. W. R. Taylor, the writer was
given the opportunity of examining sections of the leaves and
stems of ericaceous plants, collected on Mount Washington by
him, including Cassiope hypnoides, Chiogenes hispidula, Ledum
groenlandicum, Loiseleuria procumbens, Rhododendron lapponi-
cum, Vaccinium uliginosum, V. vitis-idaea.

The writer wishes to state here her deep appreciation of the
assistance given and the constant interest shown by Dr. J. M.
Macfarlane in the preparation of this paper.

The Underground Root and Stem Systems

The mature underground system in C. umbellata consists of
horizontal, thick, white runners or rhizomes bearing scales, in
the axil of which occur two buds, the first developing into an-
other branch or runner, the second into an adventitious root.
The end of the runner finally pushes above the surface of the
ground and produces a whorl of leaves. The roots are very
small, wiry, and do not branch profusely. This method of
underground stem branching is characteristic of a number of
typical Ericaceae. Warming (84) describes and figures it for
Andromeda polifolia L., Vaccinium myrtillus L., V. uliginosum L.,
V. vitis-idaea L., V. oxycoccus L. (figs. 19, 27, 29, 32, 35).

The underground systems of C. maculata, Pyrola rotundifolia,
P. elliptica, P. secunda, P. minor, P. chlorantha are similar to
that of C. umbellata (Warming (84) figs. 38, 39).

In P. aphylla, it consists, according to Holm (30) of runners
as in the other Pyrolas, one of which rising to the surface may

52 Henderson — Comparative Study of Pyrolaceae and

produce green leaves, another an inflorescence with only scale
leaves. Holm and Drude (12) report the presence of adven-
titious buds arising from the roots and forming flowering or
vegetative shoots. Irmisch (30) reports this same reproduc-
tion by root shoots in P. secunda and P. chlorantha. Holm
states that this also occurs in P. picta, C. umbellata, and C.
maculata, though none of the writer's material of the latter
genus shows this.

This reproduction by adventitious buds from the roots in
P. aphylla forms a connecting link with the condition in Moneses
uniflora. Warming (85) and Irmisch (32) state that in this spe-
cies there is a horizontal root from which arises a leaf-bearing
shoot terminating in an inflorescence. Warming states that
a root arises from the horizontal one at the base of each shoot.
In other words, there is no rhizome present; the horizontal
root here takes over the function of the rhizome in producing
leaf- and flower-bearing shoots. The occasional production of
adventitious buds on the roots of P. aphylla has become habitual
in Moneses uniflora — this being the only method of vegetative
reproduction in the latter. The main root of M. uniflora re-
sembles the rhizomes in the Pyrolas excepting for the absence
of scales on it; normal secondary roots are produced from the
root at the base of the vegetative shoot and at irregular inter-
vals along its course.

MacDougal (48) states that " Pterospora andromedea is fur-
nished with an ovoid mass of dark brown club-shaped roots
which ramify densely through a space of not more than 150-200
cc. in which the roots occupy a much greater proportion of the
volume than the included humus." There are several primary
roots which branch and rebranch, all intertwining to form a
"compact globoid mass." He states that inflorescence buds
arise from the horizontal root.

Oliver (58) has described the appearance of the root system
of Sarcodes sanguinea. He states: "The roots are attached
in great quantities to the bases of the flowering shoots and
form large and intricately woven masses of 'coralline' appear-
ance." The main axis bears numerous secondary roots which
in turn produce tertiary ones. The surface is of a "deep brown
color," showing a "certain roughness of texture" due to their
being invested with a fungal sheath. He states that the plant
is vegetatively reproduced by buds from the roots.

Monotropaceae with Reference to Ericaceae 53

In Monotropa there is a similar mass of roots. The main
root is thick, horizontal, giving rise to inflorescence buds and
to secondary roots, which are short, fleshy, and intertwine
closely to form a compact mass. Drude (11), Kamienski (39),
Queva (62) call this horizontal structure a root. Peklo (59)
calls it a "Rhizomaste." The structure is so simplified that
it is difficult to determine whether it is a root or a rhizome.

There is therefore in the underground root and stem system
a gradual condensation and simplification from genera like
Chimaphila and Pyrola, which have, as in certain typical Erica-
ceae, extensively branching rhizomes from which are produced
the vegetative axes as branches in the axils of scales, to Moneses,
where the rhizome is completely lacking and the vegetative
axes arise endogenously from an old horizontal root which
closely resembles a rhizome in appearance; to Pterospora, Sar-
codes, and Monotropa, where the horizontal root has become
very much condensed, thickened, and fleshy, also producing
vegetative buds endogenously. The roots, from being thin,
wiry, and sparsely branching in Chimaphila and Pyrola, have
the primary root enlarged and thickened in Moneses, Pterospora,
Sarcodes, and Monotropa. The secondary roots begin to show
a slight swelling at the tip in P. rotundifolia and P. elliptica.
This swelling becomes greater, secondary roots become short-
ened, thickened, and more fleshy and much more numerous,
so that there is a close compact mass of roots in Pterospora,
Sarcodes, and Monotropa.

The Root

In Chimaphila umbellata, sections of the root tip (fig. 1, 1)
show a short root cap of 4-5 layers at the very tip. This to-
gether with the epidermis arises from a common tissue, there
being only three regions of growth, the calyptro-dermatogen,
the periblem, and the plerome. The epidermal cells at the tip
are small, indistinguishable in size from the others, but soon
show gradual increase in size, particularly in a radial direction —
the radial width becoming three to four times the length. This
gradual increase occurs under the root cap, this region never
being infested by fungi. Above the root cap, the cells suddenly
become larger and square in section. These cells are infested

54 Henderson — Comparative Study of Pyrolaceae and

by septate hyphae which appear as balled-together masses in
the epidermal cells. All cells and all roots are not equally
infested — some roots show no trace of hyphae; others show a
few epidermal cells with two or three hyphal threads; others
show some epidermal cells packed full of masses of hyphae
and a hypertrophied nucleus, and cells in the same root with
normal nucleus and cytoplasm. Rommel (65) states that there
are no hyphae present in the roots of C. umbellata and that root
hairs are present. The writer has examined material from
many localities, and has found some roots of each plant infested
with hyphae and has never found any traces of root hairs.
Beneath the epidermis are one to two layers of elongated cor-
tical cells crowded with typical aggregate starch grains com-
posed of two to six simple grains, each with a distinct hilum.
These appear in safranin-stained sections as hyaline grains with
red staining hila. In the same or neighboring cells these grains
appear slightly yellowish, but still with a distinct hilum, then
yellowish brown with no hilum apparent; and finally the indi-
vidual grains merge together to form a sac of brownish material.
The secondary roots are produced endogenously.

The root tip of C. maculata has much the same appearance
as that of C. umbellata, except that the epidermal cells of the
former beneath the root cap are longer compared with their
radial width; these above the root cap are larger in comparison
with the other cells than are those in C. umbellata, are more
often, and, to a greater extent, infested with hyphae (Fig. 2, 1);
and also in those cells not filled with hyphae there appear one
to three large bladders, the walls of which are light yellowish
in color. These may be the remains of the nuclear membrane
after the nucleus has been completely destroyed by the fungus,
or enlargements formed by the hyphae similar to those described
by Groom (28) in the mediocortex of the absorbing organ of
Thismia aseroe and by Gallaud (22) in the roots of Colchicum
autumnale. The writer however did not find any hyphae
attached to these bladders.

In Pyrola rotundifolia (Fig. 1, 2) the root cap is strongly
reduced, there being only one to two layers of cells. The aspect
of the cells at the tip, under the root cap region, is much the
same as in the two species just described; but above the root
cap the epidermal cells swell out enormously, much more than

Monotropaceae with Reference to Ericaceae

55

Fig. I. Longitudinal sections (X250) of root-tips of

1. Chimaphila umbellata

2. Pyrola rotundifolia

3. Monotropa hypopitys

4. Monotropa uniflora

C = root cap, E = epidermis, F = free hyphal filaments, S = hyphal

sheath.

56 Henderson — Comparative Study of Pyrolaceae and

in C. umbellate, or C. maculate, and practically every cell be-
comes filled with balled-together masses of hyphae. There
also occurs on the outside of the epidermis a network of inter-
twining hyphae (Fig. 2, 2), forming a sheath 5 from which
extend separate filaments F which apparently take the place
of root hairs in supplying the root with water. There is a
connection between these outer hyphae and those in the epi-
dermal cells. These hyphae grow between and around the
epidermal cells, until they surround them on all sides, except
the interior; so that a tangential section shows epidermal cells

«mm^?

^AV

Fig. 2. Longitudinal sections (X300) of epidermal cells of root tip.

1. C. maculata

2. P. rotundifolia

F = free hyphal filaments, S = sheath.

separated by a pseudoparenchyma, much like that seen in
Monotropa. There is apparently no invasion of any layer
beneath the epidermis. Kramar (44) has described all stages
of the growth of the mycorhiza. The walls of the epidermal
cells first become infested with hyphae forming a pseudoparen-
chyma between the cells. Later, when the epidermal cells
become full size the hyphae penetrate the cell wall, make direct
for the nucleus and begin to form a ball around it. The nucleus
finally becomes hypertrophied and lifeless. When this occurs
and the cell is packed full of hyphae they penetrate the cell
wall again and spread out over the surface. Before this occurs

Monotropaceae with Reference to Ericaceae 57

Kramar says that the fungus is only a parasite drawing its
nourishment from the plant. After the outer covering is formed,
however, the fungus can then take the place of root hairs and
absorb water for the plant. It may also give a part of its own
assimilated food to the plant. Frank observes that as the
fungus dies each year the plant could then absorb its protein
content. Kramar also describes the contents of the subepi-
dermal cells, that they are coarsely granular and that these
cells represent a storage place where the nutritive material,
taken in by the mycelium, lies until ready to be transferred to
where it is to be used. He probably saw the starch grains
similar in appearance to those of Chimaphila when they had
become disintegrated to the extent of not showing a hilum.

Pyrola elliptica has practically the same appearance as P.
rotundifolia. P. secunda is reported to have fungal hyphae
by Irmisch (32), Rommel (65), Andres (2), Petersen (60);
P. minor by Kramar (44) and Petersen (60).

P. aphylla is described by Holm (30) as having the root free
from hyphae. This seems rather improbable in view of the
fact that all the other members of the genus have been reported
to have hyphae in the roots. The writer has not been able to
get fresh material of this for examination.

Moneses uniflora is reported by Irmisch (32) to have fungal
hyphae in the roots.

In Monotropa hypopitys (Fig. 1 , 3) Kamienski (39) and Drude
(11) describe the root tip as having one to two layers of root
cap, Drude claiming that in one variety — hirsuta — there are
two layers, and in glabra only one. Kamienski, however, says
that the number of layers may vary in the same individual.
The material examined shows one to two layers of root cap, the
outer layer crushed and flattened. All the cells at the tip are
very much alike, differentiation into plerome and periblem
occurring some distance back from the tip. Kamienski (39)
and Peklo (59) have given complete descriptions and accurate
figures of the root tip of M. hypopitys. In the Chimaphila and
Pyrola species described the fungus does not seem to invade the
root cap region either in the root cap itself, over its surface,
or in the epidermal cells beneath. In M. hypopitys, though the
epidermal cells under the root cap and the root cap cells are
not infested, the mycelium extends over the surface of the root

58 Henderson— Comparative Study of Pyrolaceae and

cap— though thinner here than over the rest of the root. As
in P. rotundifolia there is a development of pseudoparenchyma
between the epidermal cells and a continuation of this on the
exterior forming a sheath much greater in thickness than in
P. rotundifolia. This outer sheath consists of two regions, the
inner composed of closely intertwined hyphae, the outer of
more loosely arranged threads that stray out into the soil.
Kamienski claims that the hyphae never penetrate the epi-
dermal cells, but that sometimes in older parts a hypha may
penetrate an epidermal cell, which it fills with a brown content.
He also says that over the apex the sheath thins out so that
there are only several isolated filaments. MacDougal and
Lloyd (50) state that the hyphae do penetrate the epidermal
cells, forming swollen vesicles, and that the root tip is completely
invested by a thin fungal sheath. Peklo figures the penetra-
tion of haustoria into the epidermal cells. He states that
haustoria are present in all infested roots of this species. The
hyphae do not completely fill the cell as in P. rotundifolia be-
cause there is present a vacuole of yellowish brown substance.
Drude calls this a pigment ; Kamienski notes its presence in dead
cells only, and states that it is tannin; Bokorny (5) that it is
tannin in the living cells. The latter, noting that there seems
to be no difference in the quantity of this material from the
youngest to the oldest epidermal cells, concludes that it cannot
have anything to do with the nutrition of the fungus but that
it serves, on account of its strongly concentrated tannin con-
tent, as a protection against the hyphae. Peklo also notes
that beside this the resistant cuticle of the layer beneath keeps
the fungus from penetrating further into the root. He claims
that there are two ecological varieties of M. hypopitys, one
living in humus, with the roots near the surface; the other in
loamy soil, the roots deep underground. In the former, he
says that hyphae are always present and indispensable to the
life of the plant; in the latter, there are no mycorhiza in a great
majority of the roots. The writer's material, however, was
collected in a loamy soil with a surface covering of humus, the
roots about 5 dm. below the surface, and all were strongly
infested with hyphae. One difference between the epidermal
cells of this species and those of P. rotundifolia is that they are
not enormously enlarged in comparison with the rest of the

Monotropaceae with Reference to Ericaceae 59

cells — in fact the epidermal cells in M. hypopitys are smaller
than those of the layer beneath. MacDougal and Lloyd (50)
report the presence of starch in the cortex of the roots of Mono-
tropa near the tip, though the material examined by the writer
failed to show this. According to Kamienski, secondary roots
arise endogenously as in the Pyrolaceae.

The root tip of M. uniflora (Fig. 1,4) is quite similar to that
of M. hypopitys, except that at the apex the regions of growth
are even less distinguishable; the pseudoparenchyma between
the walls appears to be better developed; and the outer layer
of the sheath formed by the intertwining hyphae does not thin
out over the apex as in M. hypopitys, but continues as a layer
of the same thickness completely around the root tip. This
has been described and figured by MacDougal and Lloyd (50)
(p. 10, PI. 11). The root cap consists of one to three
(MacDougal and Lloyd 1-4) layers of flattened cells filled with
deeply staining material. The cells appeared crushed in all of
the root tips examined. The area which the root cap covers is
exceedingly limited. MacDougal and Lloyd have figured all
stages of the penetration and development of the fungus in the
epidermal cells. The nucleus becomes deformed; the hyphae
form grape-like clusters, which they consider to be atrophied
reproductive branches. They also state that starch grains
occur in the outermost layer of the cortex and that secondary
roots arise from the third layer of the cortex.

Oliver (58) has described and figured the root tip of Sarcodes
sanguinea Torr. The appearance is very like that of Monotropa
uniflora. In Sarcodes the root cap is better developed. Oliver
figures five layers. The fungus surrounds each epidermal cell,
and forms a sheath of hyphae on the surface. He states that
the hyphae never penetrate the epidermal cells, but he also
mentions that the nuclei of the epidermal cells are modified into
rod-like structures. MacDougal and Lloyd mention the pres-
ence of the mantle of mycelium that extends completely around
the root tip; that this is composed of two regions as in Mono-
tropa and that the hyphae penetrate the epidermal cells. In
the layer beneath the epidermis Oliver figures a few starch
grains. He reports that secondary roots arise exogenously.

MacDougal (48) in "Symbiotic Saprophytism" and Mac-
Dougal and Lloyd (50) have described the root tip of Ptero-

60 Henderson — Comparative Study of Pyrolaceae and

spora. It is coated with a dense brownish septate mycelium
which pushes in between the epidermal cells, penetrates them,
forming irregular vesicles and distorted nuclei, and even enters
the sub-epidermal layers. A root cap is present, and MacDougal
states that this resembles that of Sarcodes, in having more than
two layers. The mycelium covers the root cap and penetrates
the older cells in free tips, but penetrates beneath the root cap
in those roots that are in the center of the clump. He states
that starch is present in the outer cortical layers and that sec-
ondary roots arise exogenously.

Thus in the root tip region we have a gradually ascending
series in the amount of fungus present from C. umbellata, with
the epidermal cells of some roots with no hyphae — other roots
with hyphae, but not in every cell; to C. maculata with a still
greater number of the epidermal cells filled with hyphae; to
P. rotundifolia and P. elliptica with all of the cells infested and
with a beginning of a sheath of intertwined hyphae around the
root tip; then in M. hypopitys an increase in the width and
extent of this sheath and a division of it into two zones — the
outer a more loosely interwoven mass of hyphae, the inner
more compact; finally in M. uniflora a still greater width of
the fungal sheath. The descriptions of the presence of hyphae
in Sarcodes and Pterospora show a great resemblance to Mono-
tropa, but not having material to examine the writer cannot
say which has the larger amount of mycelial investment. In
Chimaphila the hyphae are probably not of much use to the
plant as the threads are only in the epidermal cells and do not
extend out over the surface. The development of an outer
sheath of hyphae in Pyrola and its great increase in amount in
Monotropa, Sarcodes, and Pterospora would allow the fungus
to be of more use to the plant and that this is true is indicated
by the lack of green coloring matter in the last three men-
tioned.

Sections both longitudinal and transverse of the oldest por-
tion of the root were examined in all of the preceding. In
C. umbellata, as the root becomes older, the hyphae penetrate
more between and into the epidermal cells, completely filling
them all. The nuclei become disintegrated, the walls thicker,
the cells die and finally fall off. The outermost layer of the
cortex may have some of its cells penetrated by the fungus,

Monotropaceae with Reference to Ericaceae 61

but this is rare. The walls of the outer cells of the cortex also
become thickened and these also fall off. Meantime the 4-5
arch bundle system has undergone secondary thickening. New
xylem cells have developed so that there is formed a solid central
cylinder of wood enclosed by a few layers of thin-walled phloem
cells. These roots are perennial and annual rings of wood
are laid down. The oldest one examined was found to have
four years of growth. The largest part of the wood seems
to be composed of pitted vessels with a few spiral ones.

C. maculata, P. rotundifolia, and P. elliptica showed prac-
tically the same appearance except that none of the mater-
ial examined was old enough to have formed annual rings
or to have had the epidermis entirely dead and sloughed off.

Fig. 3. Transverse section root of Moneses uniflora showing central xylem
tissue X 300.

In Moneses uniflora the epidermal cells become filled with
hyphae and fall off. Hyphae may penetrate into the outer
layers of the cortex. The wood is at first diarch — a few remain-
ing cells seem to indicate that this is a reduction from a tetrarch
condition (Fig. 3). Secondary wood is formed.

MacDougal and Lloyd (50) have described the structure of
the old root of Sar codes and Pterospora. The epidermis becomes
completely filled by the fungus and falls off. After this the

62 Henderson — Comparative Study of Pyrolaceae and

subepidermal cells divide radially to form a wider cortex. ' The
central cylinder of Sarcodes and Pterospora is least reduced
and its development may be traced in them with some certainty.
Here the 5-6 xylem bundles alternate with the simple phloem,
enclosing a well marked medulla. The first step in the sec-
ondary growth is the sclerotization of the medulla, and is fol-
lowed by the lignification of this tissue in Sarcodes. Next the
phloem gives rise to a cambium which develops wood inter-
nally, and bast on the outside. The latter consists for the
greater part of elongated elements of narrow lumen which do
not undergo any marked thickening of the walls. The wood
formed by the cambium joins directly on to the lignified medulla.
The inner ends of the primary medullary rays also undergo
sclerotization to some extent, but the outer portions show as
broad bands one or two layers in thickness with the character-
istic appearance of being compressed tangentially. The advance
of the cambium is at first fairly regular, as the cambium zone
moves outwardly beyond the first ring formed, the transforma-
tion into vessels is accomplished with such disturbance or vari-
ance from the customary manner that it is not possible to draw
a line separating the two regions. Furthermore, some of the
cambium cells of great size remain as great thin-walled elements
in the wood, or these may be arranged in radial lines simulating
tertiary rays. A region of cambiform elements, four to six
layers in thickness, may be seen entirely surrounding the xylem.
The structure formed by this behavior of the cambium resembles
that of a stem, and indicates that the roots of the two genera
in question may attain an age of two years or more."

In Monotropa hypopitys, the oldest part of the root is the
horizontal portion that gives off secondary roots and buds,
which lengthen to form flowering shoots. The structure of
this root has been described by Drude (11), Kamienski (39),
and Queva (62). The outer epidermal cells become completely
filled with hyphae and fall off. The cortex consists of 8-10
layers of cells that are much larger than the epidermal ones.
The endodermis is composed of small cells with thickenings
on the radial walls in the younger roots, but is indistinguishable
in the older ones. The fibrovascular cylinder is composed
mainly of thin-walled cells, there being only 3-4 patches of
xylem (Kamienski). Material that the writer examined showed

Monotropaceae with Reference to Ericaceae 63

four small areas of xylem. These patches consist only of 1-3
cells, generally reticulate tracheids, no vessels being present
according to Kamienski. Drude reports the presence of spiral
vessels. The writer has not seen any of the latter. Within
the small circle formed by these four xylem elements occurs
a parenchymatous mass of pith. Between and outside of the
xylem patches is a comparatively large amount of phloem,
composed of thin-walled cells, a few sieve tubes with slightly
thick walls being present. These sieve tubes are described by
Kamienski as having no true sieve plates, but that these are
replaced by transverse partitions, and that the walls are thin
at certain points. They are easily distinguished by their more
granular contents. Secondary thickening occurs; Kamienski
states that no cambium is present, and that the secondary
tracheids are disposed either singly or in groups toward the
pith or sometimes away from the pith. The phloem also grad-
ually increases in the number of cells, until they join to form a
ring around the wood. Queva (62), on the other hand, de-
scribes the secondary wood as being formed only in a centrif-
ugal direction — outside of the primary tracheids — and that
after this, cambial arcs form and produce a continuous ring of
xylem and phloem. Older roots examined by the writer showed
an increase in the number of xylem elements and between the
xylem and the phloem a layer of thin-walled cells resembling
a cambium. The roots were not old enough to show the con-
tinuous ring of xylem and phloem described by Queva.

No complete description of the old root of M. uniflora has
been found by the writer. The epidermal cells become filled
with hyphae, but do not appear to drop off as soon as in M.
hypopitys. The mycelial sheath still surrounds the oldest part
of the root. The fibrovascular cylinder has a very irregular
distribution of elements. The 3-4 single tracheids in young
roots have increased in number, until there is quite a large area
of wood, mostly reticulated tracheids. The secondary growth
in thickness, from the material examined, seems to occur in
the manner described by Queva for M. hypopitys.

In the least saprophytic ones the epidermis soon dies and
falls off, carrying with it the fungal hyphae as in Chimaphila
and Pyrola. In Monotropa and especially M. uniflora the
epidermis is still living and filled with hyphae when the root

64 Henderson — Comparative Study of Pyrolaceae and

is quite old. The less saprophytic ones have a fair amount
of wood developed and a comparatively small amount of phloem
as seen in Chimaphila and Pyrola. In Sarcodes, Pterospora,
and Monotropa the wood is less developed, the amount of phloem
being considerably greater. This greater production of phloem
and reduction in the amount of wood is characteristic of sapro-
phytic plants.

The Rhizome

Transverse and longitudinal sections of the underground
rhizome of Chimaphila umbellata were examined. On the ex-
terior is an epidermis composed of somewhat rectangular cells
(on transverse section) with rounded angles and thick walls.
The outer wall is much thicker than the others and has on its
exterior a layer of ridged cuticle. Interior to this is the cor-
tex, composed of 7-9 layers of rounded thin-walled cells. The
walls of the outer two layers of the cortex and those of the
epidermis become thickened, forming a cork-like region, so
that in cutting sections these all split off together. The cor-
tical cells gradually increase in size toward the interior. The
innermost layer, the endodermis, is composed of narrow rect-
angular, but somewhat irregular, cells. These are thin-walled
and show in section four to five cells filled with tannin. Internal
to this is the fibrovascular system with an external small area of
phloem and much larger area of wood. The pitted vessels ap-
pear square on transverse section and there are numerous uni-
seriate medullary rays through the wood. In C. umbellata there
were found as many as four annual rings, an evidence that the
rhizome is perennial. Inside the wood is a cylinder of rounded
thin-walled cells, the pith; starch grains occurring as either single
or aggregate clusters are numerous in the epidermis, cortex,
and pith.

The structure of the rhizomes of C. maculata, P. rotundifolia,
P. elliptica is very similar to that of C. umbellata. Conglomer-
ate crystals are present in P. rotundifolia and P. elliptica but
absent in C. umbellata and C. maculata. There is no rhizome
in Moneses uniflora or Monotropa.

The Ascending Axis
Transverse sections of all the species of Pyrolaceae and Mono-
tropaceae investigated were taken at the base of the ascending
axis just below the lowest set of scale leaves.

Monotropaceae with Reference to Ericaceae 65

In Chimaphila umbellata, the outline of the section is rounded
pentagonal. The epidermis is papillate, the cells having a
thick cuticle covered with wax. The epidermal cells and four
layers of the cortex beneath have rather heavy thickened walls,
so that this appears as a corky tissue. Inside this region are
about four more layers of rounded, thin-walled cortical cells,
and then the endodermis, a layer of narrow rectangular cells
with the radial walls thickened. Within this is a narrow zone
of phloem and a much wider one of wood. The material exam-
ined shows three annual rings. There are five groups of primary
bundles toward the interior. Innermost is the pith composed
of thin-walled rounded cells. These cells and those of the
cortex contain numerous conglomerate crystals and aggregate
starch grains.

In C. maculata, the sections resemble closely those of C.
umbellata, except that the outline is rounded triangular; there
are not as many thick- walled layers of cortex; there are only
three groups of primary bundles present; and crystals are not
as numerous as in C. umbellata.

In P. rotundifolia, the outline is rounded triangular; the
epidermis is not papillate, the cells being only slightly curved
outward ; only the two outer layers of the cortex are thickened ;
there are 7-9 layers of unmodified cortex; crystals appear to
be present in the cortex only. There are no distinguishable
groups of primary bundles; otherwise the section appears like
that of C. umbellata.

In P. elliptica, the structure is almost exactly similar to that
of P. rotundifolia.

The axis of Moneses uniflora is not as thick and woody as
those preceding. The outline is circular. The epidermal cells
are only slightly curved outward. They and the outermost
layer of the cortex are only slightly thickened but not nearly
as much as in Pyrola and Chimaphila. There are about six
layers of cortex in all. The endodermis resembles those of
the preceding. There is a small amount of phloem, one year's
growth of wood with four groups of primary bundles toward the
interior, and a small area of pith. Single and aggregate starch
grains are found in the cortex and pith, but there appear to be
no crystals present.

The structure in Monotropa hypopitys is greatly simplified.
The section is circular and wider in diameter than in any of the

66 Henderson — Comparative Study of Pyrolaceae and

preceding, as in Monotropa the axis becomes fleshy. The epi-
dermal cells are not papillate, and have a cuticular and waxy
covering which is not as thick as in Chimaphila. There is no
outer thick-walled area of cortex. It is composed of about
ten rows of large thin-walled hexagonal cells, some of which
have slightly thickened walls and contain tannin. The inner-
most layer of the cortex consists of narrow elongated cells —
the endodermis. The central vascular cylinder consists of
several bundles separated from each other by medullary rays.
Each consists externally of a wide patch of phloem — equal in
width to that of the wood. The wood consists of 10-15 ce^s
in each group. Those according to Kamienski (39) are tracheids
with either spiral or annular thickening. This ring of wood
consists of both primary and secondary wood.

The structure in M. uniflora is exactly similar to that of
M. hypopitys.

Sections of the base of the ascending axis of Pleuricospora
fimbriolata were taken from boiled up herbarium material. The
cortex is wider than in Monotropa. The vascular cylinder is
more united than in the others, there being a continuous ring
of phloem and an almost solid interior of wood — the pith being
very small in amount. The wood is much greater in amount
than in Monotropa.

Epidermis

The epidermis was examined at the middle of the ascending
axis or flowerstalk. In C. umbellata, it is composed of rect-
angular cells with thick walls and with an exterior covering
of ridged cuticle. Chloroplasts occur in the cells. Small
papillae, also with ridged cuticle, are present on some of the
cells. Stomata are present, but not numerous. They are
small and appear normal and functional; some are of the nor-
mal type with the slit parallel to the longitudinal axis — others
show all stages of twisting of the guard cells until the opening
between them becomes exactly transverse. The series is the
same as in M. uniflora except that in the latter the guard cells
are greatly enlarged and distorted. These transverse stomata
are present on the stems of Viscum album, Arceuthobium, Anti-
daphne, Loranthus, Lepidoceras, Nuytsia of the Loranthaceae;
Choretrum, Mida, Myoschilus, Anthobolus, Santalum album,

Monotropaceae with Reference to Ericaceae 67

Thesium of the Santalaceae; Cassytha of the Lauraceae; Sali-
cornia, Casuarina, Staphylea pinnata. The first three groups
are all, with the exception of Nuytsia, parasitic, so that the
presence of transverse stomata on the stem may indicate an
abnormal system of nutrition.

In C. maculata the epidermis is almost exactly similar to that
of C. umbellata, except that the papillae seem slightly longer
and have somewhat thicker walls.

In P. rotundifolia the epidermal cells are longer and thinner
walled; there are no papillae; stomata are more numerous and
are mostly of the longitudinal slit type.

The epidermis of P. elliptica is similar to that of P. rotundi-
folia, except that stomata are very rare and in all pieces of
epidermis examined were of the longitudinal type.

In Moneses uniflora there are papillae present as in Chima-
phila. The epidermal cells are long, narrow, and thin-walled,
and contain chloroplasts as in the preceding. In all the mate-
rial examined the writer found no stomata present.

In Sar codes sanguinea the epidermal cells are also long, nar-
row, and thin-walled. Oliver states that no stomata are present
on the flower stalk (the ascending axis), but the writer finds
that, though rare and somewhat distorted, stomata are un-
doubtedly present. Simple stomata with a longitudinal slit as
in M. hypopitys (Fig. 4, 4) occur. Others have the guard cells
twisted around so that the slit is diagonal (Fig. 4, 2). Others
are remarkable in that they have three guard cells (Fig. 4, 3).
At the base of the axis a few glandular hairs are present. These
increase in number toward the top of the flower stalk. They
are multicellular with a thick stalk and a slightly rounded head.

In Pterospora andromedea the epidermal cells are similar to
those of Sarcodes. This is the only member of the Monotro-
paceae up to the present that has been described as possessing
stomata on the flower stalks. MacDougal (48), p. 38, states
"Pterospora is the only dicotyledonous plant without chloro-
phyll beside Cotylanthera that is furnished with stomata."
These stomata are of the normal type with the slit parallel
to the longitudinal axis. They are very rare. Hairs of two
types occur on the flowerstalk; simple uniseriate hairs com-
posed of 3-4 elongated cells the last one club-shaped ; glandular
hairs with a multicellular stalk and a multicellular club-shaped
head.

68 Henderson — Comparative Study of Pyrolaceae and

In Monotropa hypopitys, the epidermal cells are similar to
those preceding. Stomata are present though rather rare.
Previous investigators are unanimous in saying that stomata
are absent on the flowerstalk (9. 48, 73). The writer finds
two types of stoma — the normal type with the slit parallel
to the longitudinal axis (Fig. 4, 4) and that with the slit trans-
verse to the longitudinal axis (Fig. 4, 5). The first type is

Fig. 4. Stomata X 290.

1. On lower epidermis scale of Monotropa uniflora
2-IO. On epidermis of ascending axis of:
2-3. Sarcodes sanguinea
4-5. Monotropa hypopitys
6-10. M. uniflora

more frequent in M. hypopitys. The guard cells are extremely
large and often much distorted. Whether or not these stomata
function one cannot say, but their presence indicates that they
were functional at one time, at least when the plant was young
and before the guard cells were pulled apart by the rapid upward
growth of the flowerstalk. Unicellular hairs with ridged cuticu-
lar thickenings are present on the epidermis.

Monotropaceae with Reference to Ericaceae

69

In M. uniflora the epidermal cells are similar to those of
M. hypopitys. The transverse type of stoma is much more
frequent, the longitudinal type being rather rare. Fig. 4, 6-10
shows a series in the formation of a transverse stoma from the
longitudinal. 6 is a normal stoma. In the growth of the stalk
the guard cells gradually become pulled apart and slightly
turned around as in 7. In 8 the slit is diagonal, 9 shows
the cells almost completely around, and finally 10 shows the
directly transverse slit.

Portions of the epidermis of boiled-up material of Pleuri-
cospora were found to have shorter thicker walled cells than the
preceding. No hairs or stoma ta have as yet been seen but the
material at hand was limited, so this does not finally preclude
the possibility of their occurrence. Hairs similar to those found
on the axis in Pyrolaceae and Monotropaceae are characteristic
of many of the Ericaceae. This is another close similarity
found in the three families.

Fig. 5. Transverse section of ascending axis, showing sector of fibrovas-
cular system X 200.

1. Chimaphila umbellata

2. Pleuricospora fimbriolata

X = xylem; P = phloem; HB = hard bast; C = cortex.

Transverse sections of the ascending axis were taken at a
point half way between the flower or flowers and the leaves,
or in the Monotropaceae half way between the flowers and the
base of the ascending axis.

In C. umbellata the epidermis appears as a hollow cylinder of
oval cells with a thick outer cuticle and covering of wax. The

70 Henderson — Comparative Study of Pyrolaceae and

papillate hairs appear on transverse section. Internal to the
epidermis are two to three layers of thick-walled, then two to
three layers of thin-walled, cells forming the cortex. The
outer 3-4 layers of the phloem are heavily lignified forming an
area equal in width to the succeeding area of thin-walled phloem.
There is only one year's wood developed, but this is rather
large in amount (Fig. 5, 1), the entire area being one and a half
times in width that of soft and hard bast together. Internal
to the wood is a rather large area of pith composed of thin-
walled rounded cells with small intercellular spaces.

In C. maculata the epidermal papillae appear more numerous
and more strongly developed than in C. umbellata. There is
an outer thick-walled area of three layers and an inner thin-
walled area of 3-4 layers of cortex. There are four layers of
thick-walled hard bast which is slightly greater in width than
the soft bast. The wood is not as strongly developed as in
C. umbellata, the width being equal to that of the soft and
hard bast combined. The pith is similar to that of C. um-
bellata.

In P. rotundifolia the epidermis is not papillate. There are
two layers of thick-walled, three of thin-walled, cortical cells.
The hard bast is five layers in thickness, being much greater
in width than that of the soft bast. The wood is developed
about as much as in C. maculata, being equal in amount to the
width of the hard and soft bast together.

In P. elliptica the appearance of the section is very similar
to that of P. rotundifolia.

In Moneses uniflora the epidermis is not papillate. All the
cortical layers (5-6) are thin-walled. There are four to five
layers of hard bast that dip in slightly between the patches of
soft bast so that the latter does not form a continuous ring
as in all the preceding. Each patch consists only of a few
small thin-walled cells. Internal to these is a ring of wood
that is very poorly developed in comparison with all those
preceding. In width it is only slightly greater than that of
the soft bast patches.

In Monotropa hypopitys the axis becomes more fleshy and
wider in diameter than in any of the preceding. There are
papillate hairs present that are more numerous and much longer
than in Chimaphila. Internal to the epidermis are about twenty

Monotropaceae with Reference to Ericaceae

7i

JC

H8

Fig. 6. T. S. ascending axis Pterospora andromedea X 200.

72 Henderson — Comparative Study of Pyrolaceae and

layers of rounded thin-walled cortical cells. There is a very
great development of hard bast (about 8-9 layers) which dips
in between the numerous phloem patches and joins with the
indurated medullary rays. It is greater in width than the soft
bast, which is well developed in this species. The latter has
a width about twice that of the wood, which is extremely re-
duced in amount, there being only about five to eight cells in
a group interior to the many celled phloem patch.

The structure in M. uniflora (Fig. 7) resembles that of M.
hypopitys except that no hairs are present on the epidermis,
and the amount of hard bast is not as great, there being only
about 4-5 layers.

■H8

Fig. 7. T. S. ascending axis Monotropa uniflora X 200.
L = lignified pith.

MacDougal (48) describes the structure of the ascending axis
of Pterospora andromedea as follows: "The epidermis is com-
posed of elongated spindle-form elements. The epidermis and
the two underlying layers are slightly lignified. The outline
is exceedingly crooked, and shows deep invaginations directly
over large air-chambers, while in other places distinct fixed
stomata are to be found. The cortex is composed of cylindrical
elements arranged in circles with intercellular spaces which
increase in size outwardly. The fibrovascular bundles contain
one or two annular and two or three spiral vessels, with a few
alongated elements of slight differentiation, which are perhaps

Monotropaceae with Reference to Ericaceae 73

tracheides. A number of closed sieve tubes are present. The
xylem bundles do not form a ring, but immediately external to
the bast-region is a complete cylinder of heavy sclerenchyma
composed of ten to fifteen layers of cells. Both the xylem and
the sclerenchyma are distinctly lignified. The medulla is from
3 to 6 mm. in diameter, making up about half of the cross sec-
tion of the stem, and is composed of cylindrical elements with
ample intercellular spaces." In the material examined by the
writer (Fig. 6) the hard bast area was found to be composed of
even more than fifteen layers of cells. The phloem was found
to be greater in width than the xylem. The xylem in Pterospora
is better developed, i. e., has more cells and more heavily thick-
ened walls, than in Monotropa, but is less so than in Pleuri-
cospora.

Sections of the axis of Pleuricospora fimbriolata were taken
from a boiled-up herbarium specimen. The axis is wider than
in Monotropa, about the same thickness as in Pterospora, and
consists mainly of cortex and pith composed of rounded thin-
walled cells. The vascular system (Fig. 5, 2) forms an irregular
ring placed nearer the exterior than the center of the section.
No hard bast is present; the area of soft bast is about equal
to that of the wood which is better developed than in any other
members of the Monotropaceae examined, there being more
cells and the walls more heavily thickened.

According to Solereder (73), the axis of Schweinitzia odorata
resembles that of Sarcodes and Pleuricospora in that it does
not possess a ring of hard bast fibers.

Sections of the axis of Sarcodes sanguinea were cut some
distance below the middle. They showed a great increase in
size over any of the others examined. The main bulk of the
axis is made up of thin-walled cells of the cortex and pith. The
bundles as in Pleuricospora form an irregular ring which in
position is nearer the exterior than the center. There is no
hard bast produced in Sarcodes. In the section examined by
the writer, the phloem was about equal in width to the xylem.
The phloem forms a continuous ring; the groups of xylem being
separated from each other by medullary rays. Oliver (58),
Fig. 49, figures three of these bundles. At the level where he
examined it, the phloem is less in extent than the xylem. He
states that the bundles undergo no secondary thickening.

74 Henderson — Comparative Study of Pyrolaceae and

In comparing Monotropa, Pterospora, Pleuricospora, and Sar-
codes, these have been considered in order according to their
relative fleshiness and size. The series, given in order of their
relative structure from the least simplified to the most simpli-
fied, taking as a basis of comparison the relative amount of
phloem and wood produced, would be Pleuricospora, with a
continuous ring of well developed wood equal in width to the
phloem, Sarcodes and Pterospora, with the wood not continuous
and not so strongly developed but equal in width to the phloem,
then Monotropa, with the wood only present in small patches
of about five to eight cells in a group and the phloem very well
developed.

Sections of the axes of the youngest shoots of some typical
ericaceous plants were examined. These resemble closely
those of Chimaphila and Pyrola. In Ledum groenlandicum,
Gaultheria procumbens, Rhododendron lapponicum, the epidermis
and one to two outer layers of the cortex are thickened just as
in Chimaphila and Pyrola. The inner cortex is less regular,
there being wider and more irregular intercellular spaces than
in Chimaphila and Pyrola. There is a ring of hard bast present
as in nearly all the Pyrolaceae and Monotropaceae. The wood
is generally wider than in Chimaphila. The pith is slightly
different, there being present trabeculae of thicker walled cells.
No comparison can be made between Chimaphila and Pyrola
on the one hand, and older sections of ericaceous plants, as
none of the former lives over four years and many of the latter
live for many years becoming very woody.

As regards the structure of the epidermis, cortex and the
development of cork tissue there is a complete series from
shrubby Ericaceae to the most simplified soft fleshy types of
Monotropaceae. In Rhododendron lapponicum no cork cam-
bium is developed until the third or fourth year when one de-
velops out of phloem tissue. This origin of the cork cambium
has been reported by Vesque (82). Layers of cork are laid
down and all of the external cortex and epidermis fall off. In
the more woody Pyrolaceae, Chimaphila and Pyrola, there is a
tendency for the outer thicker-walled layers of the cortex to
separate from the rest during the second and following years.
As the plant only lives for 3-4 years no opportunity for the
formation of a cork cambium takes place. In the saprophytes

Monotropaceae with Reference to Ericaceae 75

which live above ground only one season, there is no separa-
tion of the cortex even suggested. Pterospora is reported by
MacDougal to have the epidermis and the two outer layers of
the cortex slightly lignified, but in the others the cells are not
at all thickened.

The structure of the axis also illustrates a gradual descend-
ing series from typical Ericaceae, with strong development of
wood, to Chimaphila, which is almost as woody as some of the
sub-shrubby Ericaceae, to Pyrola, which is less woody, to Mon-
eses, with still less wood developed, then with a still decreasing
amount of wood development through Pleuricospora, Sarcodes,
Pterospora, and finally to Monotropa, where the wood is ex-
tremely small in quantity and the phloem well developed.

The Leaf

In C. umbellata the stem produces above ground alternating
groups of 2-3 scale leaves and whorls of 3-5 foliage leaves.
The leaves persist for four years. The last whorl of leaves
and the inflorescence appear in the same year so that no scale
leaves occur between the youngest whorl of leaves and the
inflorescence. The leaves are leathery, somewhat narrow at
the base, becoming wider above with a serrate margin, dark
green above, lighter below.

In C. maculata, there is the same alternating arrangement
of scales and leaves as in C. umbellata, except that only three or
four foliage leaves are produced in a whorl. The inflorescence
does not appear in the same year as the last whorl of leaves as in
C. umbellata, so that above these at the base of the flowers talk
three scales are borne. These are the scales which protected
the inflorescence over the preceding winter. The leaves are
leathery, lanceolate, with an acute apex, serrate, dark green
above (though lighter than in C. umbellata), with white spots
especially along its midrib, and rather reddish green beneath.

In P. rotundifolia, the arrangement of scales and leaves is
like that in Chimaphila, except that one to three foliage leaves
may be produced in one year, and that these are not arranged
in a whorl as in Chimaphila. The leaves are evergreen, lasting
from two to four years, and the node between two successive
years leaves is short, so that the leaves of two to three years
appear to be in a common rosette. They have long, narrowly

76 Henderson — Comparative Study of Pyrolaceae and

winged petioles, and large orbicular crenulate, prominently
veined, dark green, leathery blades.

In P. elliptica, there occurs the same succession of scale
leaves and foliage leaves, ending with three scale leaves at
the base of the flowerstalk, this arrangement being character-
istic of the genus. The leaves are evergreen, lasting from two
to three years. They have narrowly winged petioles, shorter
than in P. rotundifolia, and large elliptic crenulate blades, which
are smaller in size, less prominently veined, lighter green, and
less leathery than in P. rotundifolia.

In P. secunda, the leaves are evergreen and smaller than in
P. elliptica. The petioles are shorter, the blades much less
leathery, and lighter green in color.

In P. minor, the leaves are evergreen and of the same size,
or smaller than in P. secunda.

In P. chlorantha, the leaves are evergreen and of the same
size or smaller than in P. minor.

In P. aphylla Holm (30) has described and figured a rosette
of small green leaves "provided with a distinct petiole and
a blade varying from lanceolate to broadly ovate, obtuse or
slightly pointed." Below these occur a few scale leaves, and
above them the bud for the next year's inflorescence. The
green leaves are deciduous and are produced in alternate years
with the flowering shoot. For this reason some have incor-
rectly described the plant as aphyllous. Lateral inflorescences,
which arise in the axil of a scale, occur, and are therefore not
preceded by green leaves.

In Moneses uniflora, the arrangement of alternating whorls
of scales and leaves is similar to that of the genus Pyrola. The
leaves, though not at all leathery, are evergreen for two to
three years. They are small elliptic, acute, with a serrate margin.

In Pterospora andromedea, the leaves are reduced to brown-
ish scales arranged closely together, in fact overlapping at the
base of the flowering axis, but becoming more distant from
each other toward the region where flowers are produced. Those
at the base of the stem are small and somewhat triangular.
Further up the axis, these become narrower and longer, and
finally they decrease in size toward the flowers. Those at the
base are smooth on the outer (lower) surface, but toward the
top of the flower axis they become more and more hairy gland-

Monotropaceae with Reference to Ericaceae 77

ular. Stalked glandular hairs occur on the margins of all
scales.

In Sarcodes sanguinea, the leaves, as described by Oliver (58),
"show a gradual transition in form from the small scales at the
base to the thick fleshy ones covering the greater part of the
vegetative portion of the shoot. These again pass over into
the linear bracts. The leaves and bracts are fringed with
stalked glands." The outer surfaces of the leaves toward the
base of the flower stalks are smooth ; toward the tip they become
slightly pubescent, the flower bracts becoming quite markedly
pubescent. All the leaves and bracts are crimson in color.

In Monotropa hypopitys, the leaves are reduced to small oval-
triangular, somewhat thick, yellow, slightly puberulent scales,
pressed tightly against the flowering axis. They are arranged
closely together at the base of the flower axis, but become
separated further up.

In M. uniflora, the leaves are like those of M. hypopitys,
except that they are white and glabrous and much more mem-
branous.

In Pleuricospora fimbriolata, the leaves are likewise reduced
to pale brownish scales that are small, imbricating at the base,
but become larger and more distant from each other above.
They all have hair-like processes on their margins, resembling
somewhat an incompletely developed stalked gland of Sarcodes.

This reduction of leaves to scales and the general arrangement
of these are similar in Allotropa, Schweinitzia, and Newberrya.

In typical Ericaceae, as in the Pyrolaceae, there is produced,
after the foliage leaves of one year, a leaf or flower bud covered
with bud scales. This gives the alternating series of scale
leaves and foliage leaves seen in the Pyrolaceae. In the Eric-
aceae the scales are deciduous, falling off shortly after the leaves
unfold. In the Pyrolaceae, they are persistent and green, or
greenish membranous for the season, then becoming brownish
and lasting for 2-4 years. This persistence of the scales through
the season also occurs in all members of the Monotropaceae.
Here they are enlarged and take the place of foliage leaves.

In Ericaceae the foliage leaves are always green and usually
quite leathery, in most also evergreen for two or more years.
The genus Chimaphila and some of the Pyrolas, i. e., P. rotundi-
folia, are also evergreen and leathery, but in other species of

78 Henderson — Comparative Study of Pyrolaceae and

Pyrola the leaves become less leathery until in P. (Moneses) uni-
flora the leaf has a characteristic deciduous structure but is still
evergreen. In P. aphylla, the foliage leaves are deciduous, and
the ascending axis, which bears the flower, produces only scale
leaves along its course. This species is a connecting link be-
tween the evergreen type and that of all the members of the
Monotropaceae, where no foliage leaves are produced, their
place being taken by expanded fleshy scale leaves.

The question arises as to whether the scale leaves in the
Monotropaceae are the homologues of the scale leaves, or of the
foliage leaves in Ericaceae and Pyrolaceae. In the writer's
opinion, they represent both scale leaves and foliage leaves.
The deciduous scale leaves of the Ericaceae have gradually
become green and persistent, then expanded, enlarged and much
more numerous. The green leaves have become smaller and
all have changed in color from green in the Ericaceae and Pyrol-
aceae to brownish in Pterospora, crimson in Sarcodes, pinkish
yellow to yellow in M. hypopitys, to pinkish white or white in
Monotropa.

The Leaf — Microscopic Structure

The microscopic structure of the leaf of C. umbellata has been
described and figured by H. E. Petersen (60). His material
was collected in Denmark, but corresponds according to his
description and figures almost exactly with the material exam-
ined by the writer. The epidermis, both upper and lower, is
heavily cuticularized, and, on the surface of this, a layer of wax
is present. (The latter is not mentioned by Petersen.) The
epidermal cells are rectangular on transverse section, wavy
walled on surface view. There is a three-layered palisade tissue
beneath this, and a spongy mesophyll composed of thin-walled,
branched cells with large intercellular spaces. Stomata are
present on the lower epidermis on the same level as the other
epidermal cells. Petersen reports the presence of starch in
both the upper and lower epidermis, and the absence of hairs
and hydathodes. Starch is very plentiful in both palisade and
spongy mesophyll, though he does not make note of this. He
also does not mention the presence of conglomerate crystals of
calcium oxalate, which the writer finds occur in large numbers
throughout the spongy mesophyll. These occur as small con-

Monotropaceae with Reference to Ericaceae 79

glomerate crystals, about one-third the size of a normal mesophyll
cell, and as much larger masses, completely filling an enlarged
mesophyll cell. They are mentioned by Rommel (65) who has
also given a short description of the microscopic structure of
the leaf similar to Petersen. Both of these descriptions, how-
ever, deal with only the laminar portion of the leaf.

At the midrib, the halves of the lamina form an angle of 900,
and the leaf extends out into a ridge. The upper epidermis is
similar to that over the lamina, consisting of rectangular cells (on
transverse section) with thickly cuticularized walls, especially
on the outer surface. Beneath the upper epidermis, the palisade
tissue thins out to only one layer of cells, that are almost iso-
diametric. Immediately below this is the midrib vascular
bundle. Within the bundle sheath is a fan-shaped area of
xylem, below which is a narrower region of phloem, composed
of a narrow area of soft, and a wider area of hard, bast. Be-
neath the bundle are two layers of spongy parenchyma cells,
which are round, somewhat thick-walled, and closely packed
together. Next to this is one layer of rounded cells, pressed
tightly against the lower epidermis. The walls are rather
heavily thickened, so that in cutting they easily break away
from the rest of the spongy mesophyll. The lower epidermal
cells, at the midrib region, become less rectangular, more iso-
diametric than in the laminar portion, the cuticle of each cell
swelling out somewhat, so that the outline of the lower epidermis
in transverse section becomes crenulate.

C. maculata, while showing essentially the same type of leaf
as C. umbellate, differs slightly in structure. The upper epi-
dermis consists of wavy walled cells, in surface section, appear-
ing rectangular in transverse section, except that the outer
face of each cell protrudes from the surface forming a short
rounded papilla. This is, however, confined to the upper epi-
dermis, the cuticle of the lower being perfectly flat as in C.
umbellata. Chloroplasts occur in both upper and lower epi-
dermis, no hairs are present, and stomata, similar in form to
those of C. umbellata, but projecting slightly from the epidermis,
occur on the lower surface. The palisade tissue consists of
but two layers of elongated cells. The spongy mesophyll has
several layers of irregularly branching cells with large inter-
cellular spaces. Both palisade and spongy mesophyll cells are

80 Henderson — Comparative Study of Pyrolaceae and

filled with starch grains; small and large conglomerate crystals
of calcium oxalate, exactly similar to those of C. umbellata,
occur in the spongy mesophyll.

At the midrib the laminar halves form an angle of 1350-
1800, the leaves being almost entirely flat, but with a ridge
on the lower epidermis, similar to that of C. umbellata. The
upper epidermis is of the same appearance as that over the
lamina. The palisade still consists of two layers, but the cells
are shorter and wider. Beneath this are one to two layers of
spongy mesophyll. The midrib bundle is very similar in ap-
pearance to that of C. umbellata. Under the bundle are three
layers of spongy mesophyll cells, closely packed together and
with slightly thickened walls similar to those of C. umbellata,
as is also the layer just next to the lower epidermis, with its
rounded thick-walled cells, closely packed together and pressed
against the lower epidermis. The cells of the latter have be-
come shorter, more isodiametric, and have their outer walls
curved outward slightly.

The microscopic structure of the leaf of P. rotundifolia has
been described by Rommel (65), and of P. rotundifolia var.
grandiflora by Petersen (60). The material examined corre-
sponds fairly well with both of these descriptions except that
the writer finds stomata, not hydathodes, on the upper epi-
dermis; chlorophyll grains in the upper epidermal cells; crystals,
both small and large, in the mesophyll (Petersen does not
mention them, and Rommel states that they are absent); and
no differentiation into palisade and spongy mesophyll. The
epidermis is covered by a thin cuticle (much less than in Chima-
phila) and a layer of wax. Both upper and lower epidermis
consist of wavy-walled cells on surface section, rectangular in
transverse section, and contain chloroplasts. Numerous sto-
mata are present on the lower epidermis, projecting slightly
from the level of the other epidermal cells. They are fairly
frequent on the upper epidermis, especially toward the margin
of the leaf. The mesophyll consists of five to six layers of
closely packed thin-walled cells. Rommel describes the meso-
phyll as having smaller intercellular spaces near the upper
epidermis, and larger intercellular spaces near the lower epi-
dermis. Petersen's figure (p. 81) also illustrates this, but the
writer's material shows no difference between the upper and

Monotropaceae with Reference to Ericaceae 81

lower mesophyll. There is, however, the distinct transparent
median non-tannin-containing layer of the mesophyll described
by Petersen for his Danish specimens of P. rotundifolia. The
mesophyll cells, beside containing starch and tannin, contain
crystals similar to those of Chimaphila. At the midrib, the
leaf becomes prominently ridged on both surfaces. The upper
epidermal cells become enlarged and have a much thicker layer
of cuticle and wax than over the rest of the lamina. The first
layer of mesophyll cells become packed closely together and
against the upper epidermis. There are four to five layers of
rounded thin-walled cells, above and below the midrib bundle.
The latter closely resembles that of Chimaphila. The lower-
most layer of mesophyll also consists of rounded, thick- walled,
closely packed cells, tightly pressed against the lower epidermis.
The latter has its cells enlarged and has a thicker cuticle than
on the laminar region.

In P. elliptica, the structure of the lamina is exactly similar
to that of P. rotundifolia, except that the cuticle with its cover-
ing of wax is much thinner and that on the upper epidermis
is slightly thicker than that on the lower epidermis. At the
midrib, the structure is again like that of P. rotundifolia, except
that the cuticle is thinner.

The structure of the leaf of P. secunda has been described
by Petersen and Rommel and figured by the former, p. 85.
The type of leaf is very similar to that of P. rotundifolia —
there is no differentiation into a palisade and a spongy meso-
phyll. The cuticle, according to Petersen, is thicker than in
P. rotundifolia (this is not true of the writer's material) and
stomata are present also on the upper epidermis. Rommel
fails to note the latter, and also states that no crystals are pres-
ent. At the midrib, the leaf is ridged above, but not as much
as in P. rotundifolia.

The structure of the leaves of P. media and P. picta has been
described by Rommel. These are similar to P. rotundifolia
for there is no palisade mesophyll. He states that crystals
are absent in these.

The leaf of P. minor has been described and figured (p. 83)
by Petersen as having an undifferentiated mesophyll like P.
rotundifolia, but Rommel states that there is one layer of palisade
tissue. The writer finds the structure of the leaf to be similar

82 Henderson — Comparative Study of Pyrolaceae and

to that of P. rotundifolia — no palisade tissue being present.
In some places, however, a few of the cells of the layer next
to the upper epidermis become slightly elongated, as if a palisade
was beginning to form, or rather as if these were the few remain-
ing cells of what was once a palisade layer. Rommel notes the
presence of crystals. Petersen mentions the presence of stomata
on the upper epidermis, a fact not stated by Rommel. The
writer finds stomata on both upper and lower epidermis, pro-
truding slightly from the level of the epidermis. At the midrib
the leaf is ridged above and below, but even less than in P.
secunda. The bundle is not as strongly developed as in P.
rotundifolia.

In P. chlorantha, Rommel states that there is one layer of
palisade mesophyll, and that crystals are present. The writer's
material agrees with this. The epidermal cells are wavy-walled
on surface view, rectangular in transverse section. Stomata
are present on the lower surface only. There are one layer of
palisade and four layers of spongy mesophyll with larger inter-
cellular spaces than in P. rotundifolia. At the midrib there is
only a slight curving upward and a small downward ridge.
The bundle is similar to that of P. minor.

In P. aphylla, Holm (30) describes the leaf as having a normal
epidermis, covered by a thick and wrinkled cuticle, stomata are
present on both upper and lower epidermis, but more numerous
on the lower epidermis. He reports the presence of two layers
of palisade tissue and a spongy mesophyll of loosely connected
cells with large intercellular spaces.

In the genus Pyrola, most of the species examined, i.e., P.
elliptica, P. secunda, P. media, P. picla, conform to the type
of P. rotundifolia with non-differentiated mesophyll. P. minor
seems to show a transition toward the formation of a palisade.
P. chlorantha shows one layer of palisade tissue, and P. aphylla
shows two layers of palisade and a typical spongy mesophyll.
The leaves of Moneses uniflora have also been described
microscopically by Petersen and Rommel and figured by the
former (p. 86). The epidermal cells — wavy -walled in surface
section, rectangular in transverse section — have only slightly
cuticularized outer walls and contain chlorophyll. Stomata
are numerous on the lower epidermis, and are slightly project-
ing. The upper layer of the mesophyll consists of slightly

Monotropaceae with Reference to Ericaceae 83

elongated cells forming a palisade of one layer. The spongy
mesophyll cells branch irregularly and have large intercellular
spaces. Starch occurs in all the mesophyll cells. Rommel
does not mention the presence of the one layer of palisade tissue.
He also states that crystals are not present, while the writer
finds that they are present. At the midrib the upper surface
of the leaf is curved slightly upward, the lower flat. Both
upper and lower epidermis are similar in appearance to that
over the lamina. The palisade cells become somewhat shorter.
One layer of rounded mesophyll cells occurs between the palisade
and the bundle. The midrib bundle is much smaller in extent
than in either Chimaphila or Pyrola, the latter bundles being
almost equal in size to the midrib. It consists of a small area
of xylem — not radiating as in the others — beneath which are
several layers of phloem. Beneath the bundle are two layers
of spongy mesophyll cells between it and the epidermis.

MacDougal (48) has described in part the structure of the
leaf of Pterospora andromedea. He describes marginal stalked
glands similar to those on the flowering axis. According to him
"The basal scales are flecked with irregular patches of yellowish
brown areas, due to the penetration of the epidermal cells by
brownish hyphae, which completely fill them and extend over
the surface of the scales in a network." In a few scales the
writer has seen ramifying hyphae over the surface, but not
penetrating the epidermal cells. The epidermis consists of
elongated narrow cells that appear narrowly oval on transverse
section. Stomata, though not numerous, are present on the
lower epidermis. Hairs are present on the lower surface toward
the sides of the scale. These are simple unicellular protuber-
ances. The lowermost scales are smooth. Further up the
stalk, the scales become more and more hairy until those sub-
tending the flowers have long stalked glandular hairs similar
to the marginal ones. These consist of a rather long multi-
cellular stalk with an oval head composed of glandular cells.
The mesophyll is composed of numerous layers of thin-walled
hexagonal cells. The transverse section is wider at the mid-
rib, the midrib bundle being larger than the others. There is
a slight curving upward at the midrib region as in Moneses
uniflora. The bundles are more reduced than in Moneses uni-
flora. There are several xylem elements present, but the main
part of the bundle consists of phloem.

84 Henderson — Comparative Study of Pyrolaceae and

The structure of the leaf of Sarcodes sanguinea, described
by Oliver (58), corresponds closely with the material examined
by the writer. Stalked glands, similar in appearance to those
on the flowering axis, appear on the margin of the leaves. The
epidermis consists of thin-walled hexagonal cells slightly longer
than broad. On transverse section, these appear as somewhat
oval cells, those of the lower epidermis being slightly smaller
than those of the upper. Both surfaces are covered with cuticle
and a thin layer of wax, that on the outer or lower surface being
slightly thicker than that on the upper. Stomata are absent.
The basal leaves have a smooth lower epidermis, no hairs being
present. Further up the flowering axis, the lower surface
shows beginnings of multicellular glands with a long multi-
cellular stalk and a club-shaped head. These become more
numerous on the bracts which subtend the flowers. The meso-
phyll is undifferentiated, except that the first layer of cells
beneath the epidermis is composed of cells similar to it. Below
this occur 17-18 layers of hexagonal, thin-walled, closely packed
cells. The transverse section of the leaf is the same throughout,
the bundles being all nearly equal in size. They are much
more reduced than in Moneses uniflora. Several xylem vessels
are present, but the main part of the bundle is made up of
phloem.

The structure of the leaves of Monotropa hypopitys has been
described by Kamienski (39). The writer's material corre-
sponds closely with his description, except for the presence of
hairs, a fact which he does not mention. His material, however,
may have been M. hypopitys var. glabra. The epidermis con-
sists of narrow elongated cells on surface view, somewhat oval
on transverse section, with the outer wall cuticularized, ridged,
and with a thin layer of wax on the outside of the cuticle; Hairs
are present, few on the upper, numerous on the lower epidermis.
They are simple, unicellular outgrowths with small, wart-like
protuberances. Stomata are absent on the upper and very rare
on the lower epidermis. Solereder (73) mentions that they
are rare on the lower surface of the leaf, but all other writers
say they are absent. The writer found three to five on a scale.
The mesophyll consists of several layers of thin-walled, hexagonal
cells with no intercellular spaces. There is no differentiation
into palisade and spongy mesophyll. Several bundles almost

Monotropaceae with Reference to Ericaceae 85

equal in size pass through the mesophyll so that at the midrib
the scale is only slightly thicker than in the laminar region.
The bundle is even more reduced in size and number of ele-
ments than in Pterospora and Sarcodes, the woody tissue con-
sisting only of one to two elements, the main part of the bundle
being composed of phloem.

The leaf of M. uniflora is exactly similar in structure to that
of M. hypopitys except that no hairs are present in M . aniflora
and stomata are even more rare than in M. hypopitys (Fig. 4, 1).
The writer found only one stoma on the lower epidermis of each
leaf.

The leaf of Pleuricospora fimbriolata is very similar in its
structure to that of Monotropa. The epidermal cells are slightly
longer than wide on surface view, and oval in transverse section.
They are covered by a thin ridged cuticle and layer of wax.
No hairs or stomata are present. There are eight rows, at the
widest part, of thin-walled, closely packed hexagonal cells.
The midrib bundle is only slightly larger than the others and is
even more reduced than in Monotropa.

From reviewing literature on the structure of the leaves of
the Ericaceae (88, 56, 60, 73) and from an examination of sec-
tions of the leaves of various members of the family, the writer
finds that in general the structure is very similar to that of
Chimaphila, which the writer considers to be the least sapro-
phytic genus of the Pyrolaceae. The Ericaceae generally agree
with Chimaphila in having a thick cuticle, often with a coating
of wax, a mesophyll differentiated into palisade and spongy
regions, and chlorophyll in the upper and lower epidermis.
Solereder reports cuticular ridges as a common occurrence in
the Ericaceae. Their presence has been indicated in all the
Pyrolaceae and Monotropaceae examined. Papillae on the
epidermis, except at the midrib, are rare in the Pyrolaceae and
Monotropaceae, only occurring in C. maculata. They are
present, however, on the lower epidermis of Rhododendron
campylocarpum, R. thomsoni, Kalmia glauca, K. latifolia. At
the midrib in typical Ericaceae, the epidermal cells often bulge
out to form papillae. This is true of Chimaphila and Pyrola
also.

Stomata in the Ericaceae and Pyrolaceae may be present on
the lower surface only, but are frequently present on both sur-

86 Henderson — Comparative Study of Pyrolaceae and

faces. They do not, except in a few cases, have any specially
formed subsidiary cells but according to Breitfeld (88) they often
extend beyond the surface so that they overlap the surrounding
epidermal cells to some extent. This arrangement is also true
of those of the Pyrolaceae and Monotropaceae (Fig. 4).

The types of hairs present in Ericaceae are varied; Breitfeld
(88, p. 329, PI. VI), Neidenzu (56, p. 141, Pis. Ill, IV, V,
VI), and Solereder (73, p. 484, 485) have given detailed de-
scriptions and numerous figures of those found in all groups
of the Ericaceae. The forms of hairs present in the Pyrolaceae
and Monotropaceae are not numerous; in fact, hairs are absent
entirely on the leaves of all the Pyrolaceae. In Pterospora
there are present on the margin of the leaves, stalked glands
composed of a multiserrate stalk and a glandular head formed
of several cells. This type, according to Solereder, is present
in Arbutus, Arctostaphylos alpina, A. tomentosa, Enkianthus,
Gauliheria myrsinites, and G. hispida. In Sarcodes stalked
glands on the margin and also on the lower surface of the upper
leaves are very similar in structure to those of Pterospora, ex-
cept that in Sarcodes the glandular head is not as enlarged as
in Pterospora. In Monotropa hypopitys another type of hair
occurs. It is unicellular, short, blunt, and the waxy covering
is somewhat warted. This simple type is common throughout
the Ericaceae.

Chlorophyll grains occur in the epidermis of many plants
belonging to the Ericaceae and the Pyrolaceae. Petersen re-
ports their presence in the upper epidermis of the leaf of Rhodo-
dendron lapponicum, the lower epidermis of Vaccinium vitis-
idaea f. pumila, upper and lower epidermis of Vaccinium oxy-
coccus, and states that Lidforss (45) reports their presence in
the upper epidermis of Ledum palustre, Loiseleuria procumbens,
Phyllodoce coerulea, Andromeda polifolia, Lyonia calyculuta, and
in both upper and lower epidermis of Arctostaphylos uva-ursi
and A. alpina. In the Pyrolaceae, Petersen reports their
presence in the lower epidermal cells of the leaf of P. rotundi-
folia var. grandiflora, and P. minor, in both upper and lower
epidermal cells of P. secunda, P. uniflora, C. umbellata, and the
writer has noted their presence in both upper and lower epi-
dermal cells of P. elliptica and C. maculata.

The palisade tissue varies in Ericaceae from five to six layers
in Rhododendron lapponicum, Vaccinium vitis-idaea and Ledum

Monolropaceae with Reference to Ericaceae 87

palnstre; to three to four in Loiseleuria procumbens, Epigaea
repens, and Lyonia calyculata; to two to three in Ledum groen-
landicum, Kalmia latifolia, Agarista revoluta, Diplycosia pilosa,
and Andromeda polifolia; to two in Cassiope hypnoides, Gay-
lussacia pinifolia, Dendrium buxifolium, and Vaccinium uligi-
nosium: to one in Vaccinium myrtillus and Chiogenes hispidula.

The Pyrolaceae show a gradual reduction in the number of
palisade layers. Chimaphila umbellata has three; C. maculata
two; P. chlorantha and Moneses uniflora one; P. rotundifolia,
P. secunda, P. minor all have no palisade tissue developed, nor
is it present in any of the Monotropaceae.

Conglomerate and single crystals are common in many mem-
bers of the Ericaceae occurring in the palisade or spongy paren-
chyma, or in both. Niedenzu (56) has given a table of their
distribution in the Arbutoideae and Vaccinioideae (pp. 175, 176).
In the Pyrolaceae only conglomerate crystals occur and these
are found in the spongy mesophyll. The writer has found them
in C. umbellata, C. maculata, P. rotundifolia, P. elliptica, Moneses
uniflora. Rommel (65) reports their presence in P. minor and
P. chlorantha. The writer has not found any crystals in the
leaves of any of the Monotropaceae examined.

The Inflorescence

In C. umbellata, the flowerstalk is about 1 dm. long with
no scales. The flowers are arranged in a 2-8-flowered corymb,
each flower on a slightly recurved pedicel 8-12 mm. in length.
This comes off in the axil of a linear subulate, smooth, deciduous
bract.

In C. maculata, the flowerstalk is about 1 dm. long with no
scales. There are 1-5 flowers arranged in a corymb; the pedicels
are pubescent and about 16-18 mm. long. The bracts are
linear, about 15 mm. long, with the margin appearing slightly
fimbriolate near the apex on microscopic examination.

In P. rotundifolia, the entire flowerstalk is 2-3 dm. in length
with two scales. There are 6-20 flowers arranged in a raceme,
each flower being borne on a smooth recurved pedicel, 6-10
mm. long, in the axil of a lanceolate bract that is 5-8 mm. long.
This also, when examined microscopically, shows a somewhat
fimbriolate margin toward the apex.

88 Henderson — Comparative Study of Pyrolaceae and

In P. elliptica, the flowerstalk is 1-2 dm. in length with 0-2
scales; the raceme is 6-15-flowered; the bracts are linear-lanceo-
late, 5-6 mm. long, about equal in length to the smooth pedicels.
The margin of the bracts is entire.

In P. secunda, the flowerstalk is 1-2 dm. in length, with 1-4
scales. The flowers, 6-17 in number, are arranged in a raceme
with the flowers all turned to one side. The bracts are subulate-
ovate, slightly shorter than the pubescent pedicels, which are
4-5 mm. in length.

In P. minor, the flowerstalk is 0.5-2 dm. in length with one
or two scales; there are 5-17 flowers borne in a raceme, each
on a pedicel equal in length to the 2-3-mm. subulate bracts.

In P. chlorantha, the flowerstalk is 0.5-2 dm. in length, with
a single scale. There are 2-8 flowers borne in a raceme. The
bracts are lanceolate, 4 mm. long, shorter than the pedicels,
which are 5-6 mm. in length.

In P. aphylla, the flowerstalk is 1-3 dm. in length, with
numerous scales extending from the base upward. There are
8-25 flowers borne in a raceme, each on a recurved pedicel about
5 mm. in length. The bracts are lanceolate, 3-5 mm. in length.

In Moneses uniflora, the flowerstalk bears one flower and is
0.5-1.3 dm. in length with one scale, similar in form to the one
bract. This is ovate, about 4 mm. in length, and, under the
microscope, is seen to have numerous unicellular hairs along
the margin. The margin becomes slightly fimbriolate toward
the apex.

In all members of the Monotropaceae examined, green leaves
are not produced, their place being taken by scales which grad-
ually pass with little change of structure into bracts subtending
the flowers.

In Monotropa hypopitys, the flowerstalk — in this case the
entire ascending axis — is 1-3 dm. tall. The flowers are numer-
ous, 3-15, each borne on a pubescent pedicel that is 3 mm. in
length. The bracts are narrow ovate, yellow, 10-12 mm. in
length, shorter than the flower. The outer surface and the
margin of these bracts are covered with unicellular hairs.
Toward the apex, the margin becomes somewhat irregularly
toothed.

In M. uniflora, there is one flower produced at the end of
the 0.5-1. 5-dm. -tall flowerstalk. The bracts, white in color,

Monotropaceae with Reference to Ericaceae 89

are narrower than the scales below, and are about 10-12 mm.
in length, shorter than the flower, and are quite numerous at
the base of the flower. They are much thinner in texture than
in M. hypopitys. No hairs are present in these. The margin
becomes somewhat irregularly toothed toward the apex.

In Sarcodes sanguinea, the flowerstalk is 1-5 dm. in length.
The flowers are numerous, borne in racemes, each on a pedicel
that is pubescent with short glandular hairs. Those of the
lower flowers are longer than those of the upper. The bracts,
crimson in color, become much narrower than the lower scales
and are longer than the flowers, 2.5-6 cm. long. The bracts
ensheathe the buds as they come above the surface of the ground.
Later they curve backward. The whole outer surface and the
margins of the bracts are covered with glandular hairs. Near
the apex, the margin appears somewhat toothed.

In Pterospora andromedea, the flowerstalk is 3-1 1 dm. in
length. The numerous flowers are borne in racemes. The
purplish brown bracts are linear, about 5 mm. in length, as
long as, or longer than, the pubescent pedicels. Numerous
glandular hairs are present on the margin and lower surface.

In Pleuricospora fimbriolata, the flowerstalk is 1-2.5 dm.
long. Numerous flowers are borne in a raceme. The brownish
bracts are smooth, 1-2 cm. long, have a fimbriolate margin.

In Schweinitzia odorata, the flowerstalk is 5-1 1 cm. long.
Flowers are borne in a dense terminal raceme. Bracts are
purple or purplish brown and about 8 mm. in length.

In Allotropa virgata, the flowerstalk is 1-5 dm. long. Numer-
ous flowers are borne in a raceme. The whitish bracts are
linear-lanceolate, 1-2.5 cm. in length, narrower than the lower
scales.

In Newberrya congesta, the flowerstalk is 1-5 dm. tall or less,
terminated by a "corymbiform glomerule"; "scales ovate brown-
ish— the upper ones narrower, all obtuse irregularly erose" (72).

In Newberrya spicata, the flowerstalk is mostly less than
1 dm. tall, terminated by a dense spike; scales oblong, brownish,
sometimes acutish erose fimbriate (72).

In typical Ericaceae, the inflorescence is generally racemose,
or condensed to a corymb or umbel as in Rhododendron, or
solitary axillary as in Kalmia hirsuta, Phyllodoce, Cassiope,
Chiogenes. In the Pyrolaceae and Monotropaceae, Pyrola,

90 Henderson — Comparative Study of Pyrolaceae and

Monotropa hypopitys, Sarcodes, Pterospora, Pleuricospora,
Schweinitzia, Allotropa have racemose inflorescence; Chimaphila,
and Newberrya corymbose; Moneses uniflora and Monotropa
uniflora solitary terminal flowers, thus forming a series parallel
to that of the typical Ericaceae.

The bracts in the Ericaceae are generally small, green, often
deciduous as in Chimaphila and Pyrola. In the Monotropaceae,
they have become much larger, and like the fleshy scale-like
leaves. There is a transition from forms like C. umbellata and
C. maculata, with no scales on the flowerstalk, to P. rotundifolia,
P. elliptica, P. minor, P. chlorantha, Moneses uniflora with 1-2
scales; P. secunda with generally 4; P. aphylla with still more
numerous scales and no, or rarely, green leaves, and finally
to Monotropa, Sarcodes, Pterospora, Pleuricospora, Schweinitzia,
Allotropa, and Newberrya, where the scales are very numerous,
fleshy, entirely replacing green leaves at the base and becoming
only slightly modified toward the flowers.

In the Ericaceae the plants live for two or more years before
flowering and the flower buds appear in the autumn of the year
preceding their expansion. This also occurs in all members of
the Pyrolaceae. In the Monotropaceae the underground part
lives for a year before sending up a flowering axis and buds do
not appear above ground until the spring, when they are ready
to expand — this of course due to the fact that the ascending
axis in the Monotropaceae is annual.

The Sepals

In Chimaphila umbellata there are five green sepals united
at the base. The lobes are rounded, about 2 mm. in length,
and appear entire. Under the microscope, however, the margin
appears slightly fimbriolate.

In C. maculata, there are five small oval sepals, united at the
base. The segments are slightly longer than in C. umbellata,
being 3 mm. in length, about one-fourth the length of the petals.
Simple unicellular hairs are present along the margin.

In P. rotundifolia there are five green sepals united at the
base. The five lobes are lanceolate acute with spreading tips
and are 3-3.5 mm. long, one-half to one-third the length of the
petals.

Monotropaceae with Reference to Ericaceae 91

In P. elliptica, the lobes of the sepals are much shorter than
in P. rotundifolia; they are triangular, ovate-acute, about 2 mm.
long, not one-fourth the length of the petals.

In P. secunda, the calyx lobes are "oval or elliptic, 1 mm.
long, rounded at the apex" (72).

In P. minor, the calyx lobes are "triangular-acute or short-
acuminate, as broad as long" (72).

In P. chlorantha, the calyx lobes are "triangular, acutish
or obtusish, about as broad as long" (72).

In P. aphylla, the calyx lobes are "ovate-triangular, acute,
as long as broad, or slightly longer, about 1.5 mm. in length"

(72).

In Moneses uniflora, the calyx lobes are ovate, obtuse, 3 mm.
long, with numerous unicellular hairs along the margin.

In Monotropa hypopitys, the sepals are distinct, not united.
They consist of 2-3 narrow bract-like structures about 6-8 mm.
long, yellow in color, with a few long unicellular hairs on the
upper surface and numerous ones on the margin and lower
epidermis.

In Monotropa uniflora, there are 2-4 white bract-like sepals,
1-1.5 cm. long, which are not united. These are glabrous,
except for a few hairs on the upper surface. The margin begins
to show a slight irregular toothing near the axis.

In Sarcodes sanguinea, there are five fleshy oblong-lanceolate
crimson sepals about 2 cm. in length. Oliver states that these
are distinct, but the writer finds that they are very slightly
united at the base, and have a fimbriolate margin and the entire
outer surface covered with glandular hairs.

In Pterospora andromedea, the sepals are brownish, united;
the lobes are lanceolate, about 4 mm. in length, and are glandu-
lar pubescent.

In Pleuricospora fimbriolata, the four whitish sepals are sep-
arate lanceolate, 8-9 mm. long, and have a fimbriate margin.

In Schweinitzia odorata, the sepals are five, purple or purplish
brown, becoming lanceolate, 8-12 mm. long, usually acute.

In Allotropa virgata, there are five broad distinct white sepals,
"orbicular ovate to rhombic ovate, 4-6 cm. long, erose" (72).

In Newberrya congesta, the sepals are two brownish distinct,
linear or nearly so (72).

In Newberrya spicata, the sepals are two, brownish-"spatulate,
erose-fimbriate" (72).

92 Henderson — Comparative Study of Pyrolaceae and

In Cheilotheca there are 3-4 oblong lanceolate sepals (31).

In the Ericaceae, the sepals are usually green or brownish
(exceptions to this occur as in Cassiope hypnoides with a red
calyx), generally united, at least at the base. In Kalmiella
and Cladothamnus pyrolaeflorus, the sepals are almost distinct.
In a few the sepals are distinct as in Epigaea repens. All of
the Pyrolaceae are similar in this characteristic to the Eric-
aceae— the sepals all being slightly united. In the Monotro-
paceae, Pterospora has a slightly united reddish brown calyx
with narrow linear to lanceolate lobes. Sarcodes has a slightly
united calyx, but the sepals are petalloid, crimson and large,
almost as long as the corolla. In all of the other genera, the
sepals are separate, 5-4-3 in number and petalloid, showing a
gradual transition from small green united sepals in Ericaceae
to large petalloid distinct ones in the most reduced members of
the Monotropaceae.

In the Ericaceae, the sepals are often hairy, the hairs similar
to those found in the Pyrolaceae and Monotropaceae. In
Cladothamnus campanulatus, for instance, simple hairs like
those of Monotropa hypopitys and stalked glandular hairs like
those of Pterospora and Sarcodes occur.

The Petals

In C. umbellata, the petals are five, separate, concave, orbicu-
lar, pinkish in color, and 5-6 mm. in length. Under the micro-
scope they show a fimbriolate margin that is ciliolate with
numerous unicellular hairs.

In C. maculata, the petals are similar to those of C. umbellata
except for the color, which is white.

In P. rotundifolia, there are five distinct, spreading, white,
concave, roundish-obovate petals, about 7 mm. long with an
entire margin.

Those of P. elliptica are similar except that they are greenish
white in color and are about 6 mm. long.

In P. secunda, there are five oblong or elliptic erect petals,
4-5 mm. in length, greenish white in color.

In P. minor, there are five white or rose-colored, orbicular,
concave, erect petals, 3-4 mm. in length.

In P. chlorantha, there are five greenish white, oval or elliptic,
petals, 5-6 mm. in length.

Monotropaceae with Reference to Ericaceae 93

In P. aphylla, there are five obovate petals, 6-8 mm. in length,
whitish or tinged with brown or green on the outside.

In Moneses uniflora, the petals are five, occasionally six,
concave, orbicular, white or pinkish, about 1 cm. in length,
with a fimbriolate margin that is only visible under the micro-
scope.

In Monotropa hypopitys, there are 5-4 yellow petals (five in
the terminal flower, four in the lateral flowers). These are
about 1 cm. in length and are narrow with a saccate base. Al-
though the petals are not united, they are erect, with margins
meeting, so that the flower appears campanulate. They are
covered on the upper and lower surfaces with numerous simple,
unicellular hairs.

In M . uniflora, there are five, occasionally six, white or pink-
ish petals that appear somewhat similar to those of M. hypopitys,
except that they are larger (1.5 cm. in length) and thinner in
texture, and only sparsely hairy on the inner surface, and en-
tirely glabrous on the outer surface.

In Sarcodes sanguinea, the petals are five in number, united,
campanulate (1-1.5 cm. in length), the five lobes are broad,
rounded, and slightly spreading. No hairs are present on either
surface.

In Pterospora andromedea, the corolla is urceolate, the five
petals are white in color and united. They are 7-8 mm. in
length, the lobes ovate to reniform, very short and recurved.

In Pleuricospora fimbriolata, there are five separate, slightly
spreading, white petals, each narrowly oval, with a fimbriolate
margin.

In Schweinitzia odorata, the corolla is pink in color. It is
campanulate, with five lobes, which are ovate, shorter than the
tube. The tube is 5-saccate at the base, as in Monotropa.

In Allotropa virgata, petals are absent.

In Newberrya congesta, the corolla is urceolate. The four
lobes are ovate, about one-third as long as the tube which is
pubescent within.

In N. spicata, the corolla is oblong campanulate, the four
lobes oblong-ovate, about one-half as long as the tube which is
pubescent within.

In Cheilotheca there are three erect, linear-oblong, yellow-red
petals, 2.5 cm. in length.

94 Henderson — Comparative Study of Pyrolaceae and

In the Pyrolaceae, all of the species have five distinct petals.
This is also true of the most primitive group of the Ericaceae,
namely, the Rhododendroideae-Ledeae where Elliottia has four,
Tripetaleia three, Cladothamnus five, Bejaria seven, and Ledum
five petals. Distinct petals also occur in several members of
the Monotropaceae, i.e., Monotropa hypopitys with 5-4, M.
uniflora 5-6, Pleuricospora 5-4, and Cheilotheca three. In other
members of the Monotropaceae the petals are united. In
Pterospora, the corolla is urceolate, resembling that of Andro-
meda. In Sarcodes and Schweinitzia it is campanulate, and in
Newberrya urceolate to campanulate. Urceolate and campanu-
late corollas are quite characteristic of many of the Ericaceae.
In the Ericaceae proper one can trace all transitions from a
flat saucer-shaped corolla with separate petals, as in Clado-
thamnus; to shallow campanulate, as in Loiseleuria procumbens;
to deep campanulate, Epigaea; to campanulate becoming slightly
irregular bilobed, as in Rhododendron; to urceolate in Andromeda
and Vaccinium; to deeply urceolate in Erica and Thibaudia.
In the Monotropaceae, all these stages do not occur; there is
quite a big gap between a corolla with distinct petals, as in
Monotropa, and an urceolate one as in Pterospora, or campanu-
late as in Sarcodes and Schweinitzia. This would point to the
view that perhaps these three arose from a higher group of the
Ericaceae, and that the others arose from the Rhododendroideae-
Ledeae with distinct petals. The very great similarity of these
to each other (particularly Sarcodes and Pterospora, the writer
having no good material of Schweinitzia) in all of their parts,
and the rather great difference in structure between them and
all other members of the Monotropaceae and Pyrolaceae give
further evidence toward this view. In Chimaphila and P.
rotundifolia and P. elliptica the corolla is slightly irregular —
one petal extends downward so as to form a resting place for
the insect. This parallels the condition in Rhododendron where
the corolla is slightly bilabiate.

The presence of hairs on the inner surface of the petals seen in
Monotropa is a characteristic of many members of the Ericaceae.

The Stamens

In all of the Pyrolaceae, the stamens, ten in number, are
arranged in the bud so that the pores of the anthers point down-
ward. When the flower opens, the anthers tilt backward, so

Monotropaceae with Reference to Ericaceae 95

that the pores point upward. This is to insure pollination.
Pollen grains occur in tetrads in all of the Pyrolaceae except
P. secunda.

In C. umbellate, the filaments have two lobes toward the base.
These bear numerous unicellular hairs along the margin. The
anthers are violet in color and attached nearer the pore bearing
end, and have two short horns opening by apical pores.

The stamens of C. maculata are similar to those of C. umbellata
except that there are present on the margins of the lobes of the
filament uniserrate hairs composed of 2-3 simple cells placed
end on end. The anthers are attached near the middle.

In P. rotundifolia and P. elliptica, the filaments are not lobed
and the horns on the anthers are very slightly developed.

In P. secunda, the anthers are oblong, opening by large pores.
No horns are present. The pollen grains are single.

In P. minor, the anthers are not horned.

In P. chlorantha, the horns on the anthers are well developed,
being about 0.5 mm. in length.

In P. aphylla, the horns are well developed, being about 1 mm.
in length.

In Moneses uniflora, the filaments are awl-shaped, the anthers
are prominently two-horned, the horns 0.5 mm. in length.

In Allotropa virgata, the stamens are described as having
slender filaments, "anthers short, lobed, unappendaged, extrorse
in the bud, introrse in anthesis, the sacs opening to near the
middle by a chink." (72). This is the only member of the Mono-
tropaceae that possesses the faculty of changing the position
of the anthers in the opening bud, thus forming a connection
between the Pyrolaceae and the Monotropaceae. Pollen grains
are simple as in all the other members of the Monotropaceae.

In Monotropa hypopitys, the filaments are long, with unicellu-
lar hairs. They are pressed closely up against the ovary ex-
tending up to the stigmatic disk, the five opposite the petals
being shorter than those opposite the sepals. The anthers
are short, kidney-shaped, with transverse dehiscence, opening
into two unequal valves.

The stamens of M. uniflora are similar to those of M. hypopitys.

In Sarcodes sanguinea, the filaments are long, slightly ex-
panded at the base, extending up slightly further than half the
length of the corolla, lying in the grooves of the ovary, and

96 Henderson — Comparative Study of Pyrolaceae and

bearing on their apices the rather long expanded anthers which
open by two oval pores at the summit, turned toward the outside.

In Pterospora, the filaments are slender, the anthers short,
each sac with an awn. The dehiscence is longitudinal.

In Pleuricospora, there are 10-8 stamens; the filaments are
long and glabrous, the anthers long and narrow, opening by a
longitudinal slit.

In Schweinitzia odorata, the anthers are short, opening by
terminal pores.

In Newberry a, the stamens are 10-8 in number; the filaments
are slender with long hairs. The anthers are oblong, erect on
the tip of the filament, opening lengthwise.

In Cheilotheca, the anthers are erect on the filaments, and
have longitudinal dehiscence (31).

In the Ericaceae, types of stamen similar to all those of the
Pyrolaceae and Monotropaceae are present. Many of them
have hairy filaments, as in Chimaphila and Monotropa. Simple
oblong anthers occur as in P. minor, P. secunda, and Sar codes
with apical pores and not horned, e. g., Kalmia glauca (12, p. 26,
Fig. 17); others with exceedingly long tubes, as in Vaccinium
vitis-idaea (12 Fig. 17) longer than in any of the Pyrolaceae
and Monotropaceae. In the Pyrolaceae and Monotropaceae
dehiscence of the anthers occurs in the same way as in the
Ericaceae. Apical porous is quite common. Longitudinal de-
hiscence also occurs in practically all groups of the Ericaceae.
In the Rhododendroideae-Ledeae, Elliottia and Cladothamnus
have longitudinal, Bejaria and Ledum apical porous dehiscence.
Transverse dehiscence occurs in the group Arbutoideae-Andro-
medae.

In the Ericaceae, the anthers may or may not be awned.
In Figs. 17 and 18, p. 26, Drude (12) has figured Erica tetralix,
Arbutus unedo, Calluna vulgaris with awned anthers, Vaccinium
vitis-idaea, Kalmia glauca, Rhododendron flavum , and Leiophyllum
buxifolium without awns. As a general rule the more primitive
members of the Ericaceae with open flat expanded flowers do
not have awned anthers, those with urceolate corollas gener-
ally do. This rule also applies to the Pyrolaceae and Mono-
tropaceae; Pterospora the only member with awned anthers, has
an urceolate corolla. The awns when touched by an insect
tilt the anthers so that the pollen is dropped out on the insect's
back.

Monotropaceae with Reference to Ericaceae 97

The pollen grains of Ericaceae always occur in tetrads. This
is true also of the Pyrolaceae except for P. secunda. This
species forms a transition to the Monotropaceae where all the
pollen grains are single.

The Pistil

In C. umbellata, the ovary is somewhat flattened-globose,
with its outer surface ridged from the anthers being pressed up
against it. The outer surface is covered with simple papillar
hairs. It is completely five-celled, with a central placenta at
the base, but the upper half is one-celled through the failure
of the bilobed parietal placentae to meet and fuse in the center.
The space between the placentae is, however, very small, so
that it appears more like a five-celled ovary than a one-celled
one. The placentae bear numerous anatropous ovules. The
style is short, top-shaped. The stigma is broad and five-crenate,
with a disc-shaped border. At the base of the ovary is a circular
disc which secretes nectar.

In C. maculata, the ovary is similar to that of C. umbellata,
the outer surface being ridged from the anthers, but no hairs
are present. The number of cells, etc., is the same as in C.
umbellata. The nectariferous disc is present, similar to that
of C. umbellata. The style, however, is slightly longer; the
five stigmatic lobes protrude slightly more from the disc.

In P. rotundifolia, the ovary is five-lobed, with glabrous
outer walls. The same transition from a five- to a one-celled
state occurs, as in Chimaphila, this being characteristic of the
family. No nectariferous disc is present. The style is long
declined, with the apex turned upward. . The stigma consists
of five narrow, erect lobes, which extend out from the top of the
style that forms a rim beneath the stigmatic lobes.

In P. elliptica, the ovary and style are similar to those of
P. rotundifolia. No nectariferous disc is present. The lobes
of the stigma are however slightly longer, and the rim at the
tip of the style greater in diameter than in P. rotundifolia.

In P. secunda the ovary is five-lobed, subglobose, with ten
small nectariferous lobes at the base. The ovary is five-celled
for most of the distance; the region where the placentae fail to
meet is limited because of the deep insertion of the style. This
is erect, straight, exserted, about 4 mm. in length. The stigma
is peltate, much broader than the style, with five diverging lobes.

98 Henderson — Comparative Study of Pyrolaceae and

In P. minor there are no nectaries at the base of the five-
lobed, incompletely 5-celled, ovary. The style is short, erect,
1 mm. in length, included in the petals.

In P. chlorantha the ovary is five-lobed, with 10 small nec-
tariferous lobes at the base. The style is long, about 7 mm.,
exserted a little beyond the corolla, thickened upwards and
declined as in P. rotundifolia and P. elliptica.

In P. aphylla the ovary is five-lobed with ten small nectari-
ferous lobes at the base. The style is erect, short, 3 mm. in
length.

In Moneses uniflora the ovary is ten-lobed. It is five-celled
for the most part because the deep insertion of the style limits
the region where the placentae fail to meet. No nectariferous
lobes are present. The style is erect, inserted rather deeply
into the ovary. It widens out toward the extremity forming
a rim that is wider than that of P. elliptica, and with a stigma
consisting of five fleshy lobes that are longer than in P. elliptica.

Fig. 8. Transverse section ovary Pleuricospora fimbriolata X 25.

1. Near base.

2. At middle.

In Monotropa hypopitys the ovary is ovoid, 10-8-lobed. As
in the Pyrolaceae it is 5-4-celled at the base and one-celled
above with 5-4 bilobed parietal placentae. At the base of the
ovary there are 10-8 downward directed spur-like processes,
which extend between the stamens and secrete nectar into the
saccate bases of the petals. The style is thick and fleshy,
longer than the ovary, it and the ovarian wall being strongly
pubescent with unicellular hairs. Toward the top, the style

Monotropaceae with Reference to Ericaceae

99

expands slightly, forming a ring, below this is a circle of hairs.
The upper surface is hollowed out into a funnel-shaped 4-5-
sided stigma.

In M. uniflora the ovary is ovoid, larger and more distinctly
ten-lobed than in M. hypopitys. At the base are developed the
ten nectaries similar to those of M. hypopitys, but larger. The
style is shorter, the stigma much less hollowed out than in
M . hypopitys, appearing more like a flat disc.

In Sarcodes sanguinea the ovary is smooth, ten-lobed, each
lobe extending between two stamens and continued down into
a nectar-secreting portion. It is five-celled at the base, becom-
ing one-celled above as in Monotropa. The style is erect, bear-
ing a five-lobed stigma.

In Pterospora the ovary is ten-lobed, depressed. It is five-
celled below and one-celled above as in all the preceding. No
nectaries are present. The style is short and broad, the stigma
peltate-capitate, slightly five-lobed.

Fig. 9. Longitudinal section flower Newberrya spicata at base X 25.
F = filament, N = nectary, P = petal, S = sepal.

In Pleuricospora the ovary is ovoid, not lobed, the epidermal
cells bearing distinct pointed papillae. It is four-celled at the
base (Fig. 8, 1) and for about one-sixth the length of the ovary,
but further up (Fig. 8, 2) the four placentae fail to meet and
do not extend far into the interior of the ovary, so that it appears
as a one-celled ovary with parietal placentae. There are no
nectaries present. The style is short and broad, the stigma
depressed capitate.

In Schweinitzia the ovary is five-lobed with no downward
directed nectaries at the base ; it is five-lobed for about half the
distance; the placentae do not extend in quite as far as in Pyrola.

ioo Henderson — Comparative Study of Pyrolaceae and

The epidermal cells of the ovarian wall are slightly papillate.
The style is short and thick, the stigma disc-like pentagonal.

In Allotropa the ovary is five-lobed. It too is five-celled
at the base, one-celled above. There are ten small slightly
down-directed nectaries at the base of the ovary.

In Neivberrya the ovary is five-lobed, five-celled at the base,
one-celled above. There are ten downward directed spur-like
nectaries at the base of the ovary (Fig. 9). The ovary and style
are pubescent with simple unicellular hairs. The stigma is
depressed capitate.

In Cheilotheca the ovary is "fusiform, one-celled, narrowed
into the short cylindric style; stigma globose conical; placentae
six, parietal, bifid, the long branches on all sides covered by
numerous ovules" (31).

In the Ericaceae a five-lobed, completely five-celled ovary is
characteristic. In the Pyrolaceae and Monotropaceae there is
every transition from a five-celled ovary with central placenta,
as in the Ericaceae, to an incompletely five-celled ovary with
the placentae deep parietal and almost meeting, as in all the
Pyrolaceae, also Sarcodes and Monotropa; to an incompletely
five-celled ovary with the placentae not so deep parietal, so
that there is a good space between them, as in Schweinitzia;
to Pleuricospora where the ovary is four-celled for only a short
distance, then one-celled above with parietal placentae that
are close to the ovary wall. Drude (12) states that all the
Pleuricosporeae are one-celled. The writer finds the condition
in Neivberrya as in the Pyrolaceae; in Pleuricospora 4-1 -celled
as above ; no material of Cheilotheca could be obtained for exam-
ination. Hooker (31) describes it as being one-celled with six
parietal bifid placentae.

In Ericaceae the ovary is often covered with hairs, peltate
glandular in Rhododendron lapponicum; glandular and setaceous
in Ledum palustre; glandular in Phyllodoce coerulea. In the
Pyrolaceae and Monotropaceae simple hairs are present in
Chimaphila umbellata, in Monotropa hypopitys; and small papillae
are present in Schweinitzia odorata and Pleuricospora fimbriolata.

The ovules in all three families are always anatropous. Corre-
lated with the great numbers of ovules produced in the Pyrola-
ceae and Monotropaceae, is the rather remarkable number of
pollen tubes developed from germinating pollen grains on the

Monotropaceae with Reference to Ericaceae

101

stigma. These tubes pass down through the stylar canals and
spread over the placental surfaces. They are most numerous
in M. uniflora.

In practically all members of the Ericaceae there is at the
base of the ovary a circular or crenately lobed nectary. In
Chimaphila this is represented by a narrow collar-like rim.
It is absent entirely in P. rotundifolia, P. elliptica, P. minor,
Moneses uniflora, Pterospora, Pleuricospora. The disc or nectary
is not continuous in any others of the Pyrolaceae or Mono-
tropaceae, being represented by ten very small swellings in
P. secunda, P. chlorantha, P. aphylla, and Sar codes sanguinea;
and by ten slightly larger and down directed lobes in Allotropa,
Schweinitzia, Newberrya, and Monotropa.

Fig. 10. Longitudinal section seed of Pyrola rotundifolia X 300.
E = embryo.

The Fruit and Seed

In the genus Chimaphila, the capsule is depressed-globose,
five-valved, splitting from the apex downward. The valves
are smooth along the edges. After the flower is pollinated,
the pedicel straightens up so that the fruit is erect. This is
true for the entire family. The seeds are small, numerous, and

102 Henderson — Comparative Study of Pyrolaceae and

consist of a thin cellular coat with an endosperm composed of
a few large cells. The embryo has no form, but consists of
several cells.

In the genus Pyrola, the capsule is five-lobed, splitting from
the base upward. The valves are cobwebby on the edges, a
distinction from the genus Chimaphila. The seeds in P. rotundi-
folia (Fig. io) and P. elliptica are similar to those of Chimaphila.

In Moneses uniflora, the capsule is five-lobed, splitting from
the base upward. The valves are smooth on the edges.

In Monotropa hypopitys, the capsules are oval, 5-4-celled,
loculicidal. The seeds have a thin loose cellular covering;
the endosperm consists of a very few large cells — much fewer
and larger than in Pyrola; and the embryo itself is reduced to
nine cells according to Koch (43), and five according to Solms-
Laubach (74).

In Monotropa uniflora, the capsule is ovoid, dehiscence as in
M . hypopitys. The seeds are similar to those of M . hypopitys
in the number of endosperm cells, but appear to be even more
reduced in the number of cells in the embryo, the writer's mate-
rial showing only two.

In Sarcodes, the capsule is spheroidal, 9-21 mm. broad, with
circumscissile dehiscence, the line of dehiscence 1— 1.5 mm.
below the base of the style. Oliver has described and figured
the seeds of this. In regard to the amount of endosperm and
number of cells in the embryo he states that it is very similar
to that of M. hypopitys.

In Pterospora, the capsule is five-lobed, five-celled, loculicidal,
the valves cohering with the columella. The seeds are very
numerous, ovoid, and are broadly winged at the apex (Drude
(12), p. 10, Fig. 6-K).

In Pleuricospora, the capsule is ovoid, one-celled.

In Schweinitzia, the capsule is ovoid, five-celled, seeds nu-
merous.

In Allotropa, the capsule is "spheroidal, 4-5 mm. broad" (72).
In seeds examined by the writer, the endosperm shows the same
number of cells as in Monotropa.

In Newberrya, the capsule is short, usually ovoid. The writer
had no ripe capsules, but found the ovary five-celled, so concludes
that the capsule is five-celled also.

Monotropaceae with Reference to Ericaceae 103

In Ericaceae, the fruit is a capsule, with septicidal or loculicidal
dehiscence, or a berry. In the Pyrolaceae and Monotropaceae
the fruit is a capsule, as in the primitive Ericaceae. It is char-
acteristic of many Ericaceae, e.g., Cassiope hypnoides, C.tetragona,
Phyllodoce coerulea, that after flowering (in species with droop-
ing flowers) the pedicels straighten out so that the fruit becomes
erect. This is characteristic of all of the Pyrolaceae and Mono-
tropaceae.

In typical Ericaceae, the seeds are small, the largest 1-2 mm.
The endosperm is well developed, the embryo distinctly formed,
a root and two cotyledons always present. In the Pyrolaceae
and Monotropaceae the seeds are smaller and more numerous.
In structure they are much reduced. The endosperm in the
Pyrolaceae consists of relatively few large cells — the embryo
of about 25-30 cells with no trace of cotyledons. In the Mono-
tropaceae the number of endosperm cells is still less and the
cells are larger, the embryo also is very small, composed of only
nine or five cells (43, 74).

Summary

In the Ericaceae the plants are shrubby or sub-shrubby.
The Pyrolaceae show a gradual reduction from sub-shrubby in
Chimaphila to herbaceous in Moneses. The Monotropaceae are
entirely herbaceous. The underground rhizome, producing ad-
ventitious buds and roots, in many of the Ericaceae, in Chima-
phila and Pyrola, gradually has its function of producing buds
and roots taken over by the root which becomes long and hori-
zontal in Moneses, becoming condensed to short and fleshy in
the Monotropaceae. There is a gradual increase in the amount
of hyphal investment in the roots from Chimaphila through
Pyrola to Monotropa, the most saprophytic, correlated with a
gradual decrease in the number of layers in the root cap. In
the structure of the ascending axis, there is a gradual decrease
in the amount of wood formed, from typical Ericaceae, with
very woody stems through Chimaphila which is as woody as
some of the smaller Ericaceae, through Pyrola and Moneses
which are less woody, to the Monotropaceae, reaching its climax
in Monotropa where the amount of wood is very limited. Cor-
related with this is a gradual increase in the amount of phloem.
There is a gradual reduction in the size and structure of the

104 Henderson — Comparative Study of Pyrolaceae and

leaves from evergreen leathery in Ericaceae, Chimaphila, and
some of the Pyrolas, to less leathery in P. chlorantha and P.
minor, to evergreen leaves but with deciduous structure in
Moneses, to scales which are brownish (Pterospora), brownish
yellow or yellow (M. hypopitys), red {Sarcodes), brownish white
(Pleuricospora) , to white (M '. uniflora) . This is correlated with
the gradual increase in size and persistence of the scales from
green or brownish in the Ericaceae to herbaceous, persistent,
but still small in the Pyrolaceae, to large fleshy and colored in
the Monotropaceae. Stoma ta are very numerous on the leaves
in the Ericaceae, become less numerous in the Pyrolaceae, very
few in the scales of Monotropa and Pterospora and absent entirely
in those of Sarcodes. All of these changes are correlated with
increasing saprophytism.

The inflorescence is a raceme or condensed to a corymb, or
solitary in each of the three families. The sepals are green,
united, in the Ericaceae and Pyrolaceae; brownish, united, in
Pterospora; red, very slightly united, in Sarcodes; becoming
separate, yellow, in M. hypopitys; and white in M. uniflora.
The petals are 5-4, separate, to united shallow campanulate,
to campanulate, to irregular campanulate, to urceolate in the
Ericaceae; there are five separate expanded, to separate cam-
panulate in the Pyrolaceae; to separate campanulate in Mono-
tropa; to united campanulate in Sarcodes; to urceolate in Ptero-
spora. Stamens are generally twice the number of petals in
the Ericaceae, and always so in the Pyrolaceae and the Mono-
tropaceae. They bear two horns in many of the Ericaceae and
Pyrolaceae. They are awned in the urceolate corolla types in
the Ericaceae and in Pterospora (which has an urceolate corolla)
of the Monotropaceae. The dehiscence is apical porous, longi-
tudinal, or transverse in the Ericaceae, apical porous in the
Pyrolaceae; and apical porous, longitudinal, or transverse in
the Monotropaceae. Pollen grains occur in tetrads in all of
the Ericaceae and all of the Pyrolaceae except P. secunda. In
P. secunda and all of the Monotropaceae they are simple.

In the Ericaceae, the ovary is completely 5-4-celled. In
the Pyrolaceae, in the upper half of the ovary the placentae fail
to meet and fuse. They almost meet, so that the ovary is
practically five-celled. In the Monotropaceae, there is a gradual
decrease in the length of the placental in-growths to Pleuri-

Monotropaceae with Reference to Ericaceae 105

cospora where the placentae are placed close against the walls
and the ovary is practically one-celled. At the base, however,
even in this most simplified form, the ovary is four-celled. At
the base of the ovary in practically all of the Ericaceae, there
is a nectariferous disc. This is represented either as a collar-
like rim in Chimaphila, or ten very small swellings as in P.
secunda, P. chlorantha, P. aphylla, and Sarcodes or ten downward
directed nectaries in Allotropa, Schweinitzia, Newberrya, and
Monotropa. The ovules in all three families are anatropous.
In the Ericaceae, the seeds are small with abundant endosperm
and a well-formed embryo; reduced in size but increased in
number in the Pyrolaceae with a less developed endosperm
with larger and fewer cells, and a formless embryo composed of
about 25-30 cells; still further reduced and more numerous in
the Monotropaceae with an endosperm consisting of a few large
cells and a very small embryo composed of 9-5 cells. The
change in the ovary from five-celled with central placentae to
one-celled (nearly so in Pleuricospora) with parietal placentae;
and the increase in number of seeds and reduction in number
of cells of the endosperm and embryo are all evidence of increas-
ing saprophytism.

Conclusions

From the preceding summary, it is seen that all of the sup-
posed differences between the Ericaceae and the Pyrolaceae
are broken down, except that the ovary is completely five-celled
in the Ericaceae, and incompletely five-celled in the Pyrolaceae.
The distinction is so slight that it seems unreasonable to use
it as a basis for separating the two families. The only dis-
tinction that holds between the Pyrolaceae and the Mono-
tropaceae is the absence of green coloring matter in the latter.
In M. hypopitys, small grains are present in the epidermis of
the scales. These are probably chromoplasts which develop
from chloroplasts by degeneration.

The Pyrolaceae and Monotropaceae therefore differ from
the Ericaceae only in their gradually increasing saprophytism
and in those characters which go hand in hand with this, i.e.,
the loss of green coloring matter, the reduction from shrubs
to herbs, reduction of leaves to scales, the ovary from five-
celled with central placentae to almost completely one-celled

106 Henderson — Comparative Study of Pyrolaceae and

with parietal placentae, the increase in the number of seeds
and the reduction in their size and in the number of cells of
the endosperm and embryo. The writer has traced such a
gradual transition in structure from the simplest group of the
Ericaceae, i.e., the Rhododendroideae-Ledeae, through the
slightly saprophytic Pyrolaceae to the completely saprophytic
Monotropaceae. The genera Pterospora and Sarcodes seem in
their characters to be different from the others, as if these two
had arisen in a different line, perhaps from some of the higher
members of the Ericaceae with campanulate or urceolate corollas.
On account of the lack of distinct differences and the numer-
ous resemblances and intergrading characters between the more
primitive Ericaceae, the Pyrolaceae, and the Monotropaceae,
the writer concludes that the last two should be considered as
sub-families under the Ericaceae.

BIBLIOGRAPHY

1. Alefeld. Ueber die Familie der Pyrolaceen. Linnaea XXVIII, 1856. 71.

2. Andres, H. Die Pirolaceen des Rheinischen Schiefergebirges der
angrenzenden Tieflandes des Rheins und des Mainzer Beckens. Verhand.
des Naturhist. Ver. der Preussischen Rheinlande Westfalens 1907-09.

3. Baillon, H. Histoire des Plantes II, 1892.

4. Bentham, G., and Hooker, J. D. Genera Plantarum, Vol. II, Part II,
1876.

5. Bokorny. Selbst Schutz der Pflanzen gegen Pilze. Biol. Cent. 1899.

6. Britton, N. L., and Brown, S. Illustrated Flora of the Northern
States and Canada, 1913 Ed.

7. Campbell, D. H. Monotropa uniflora as a Subject for Demonstrating
the Embryo-sac. Bot. Gaz. 1889, 83.

8. Chatin, G. A. Anatomie comparee des vegetaux. Plantes aquatiques
et parasites, Paris, 1 856-1 865.

9. DeBary, A. Comparative Anatomy of Phanerogams and Ferns, 1884.

10. De Candolle, A. Prodromus, VII, 1839.

11. Drude, O. Die Biologie von Monotropa Hypopitys und Neottia
Nidus avis L, etc. Gottingen 1873.

12. Drude, O. Pirolaceae, Ericaceae in Engler and Prantl — Die Natiir-
lichen Pflanzenfamilien, IV, I, 1889.

13. Drude, O. Deutschland's Pflanzengeographie, I, 1896.

14. Dufrenoy, J. The Endotrophic Mycorhiza of Ericaceae. New
Phytologist, Vol. XVI, 1917, 222-228.

15. Engelmann, G. Botanical Works, 1887.

16. Engler, A. Burmanniaceae. Engler and Prantl — Die Natiirlichen
Pflanzenfamilien, II, 6.

17. Engler, A. Loranthaceae. Engler and Prantl — Die Natiirlichen
Pflanzenfamilien, III, I.

Monotropaceae with Reference to Ericaceae 107

18. Frank, B. Neue Mittheilungen uber die Mycorhiza der Baume und
der Monotropa hypopitys. Ber. d.deuts. bot.Gesellsch. Generalversammlung,
1885, XXVII-XXX.

19. Frank, B. Ueber neue Mycorhiza Formen. Ber. d. deutsch bot.
Gesellsch., Bd. 8, 1887.

20. Frank, B. Ueber die auf Verdauung von Pilzen abzielende Symbiose.
Ber. d. deuts. bot. Gesellsch., IX, 1891, 244.

21. Fraysse, A. Contribution a la biologie des plantes phanerogams
parasites. Montpellier, 1906.

22. Gallaud, I Etudes sur les Mycorhizes Endotrophes. Revue Generate
de Botanique, T. XVII.

23. Goebel, K. Organography of Plants, 1900-1905.

24. Gray, Asa. Manual of the Botany of the Northern United States,
1st Ed., 1848.

25. Gray, Asa. New Manual of Botany, 7th Ed., 1908.

26. Gray, Asa. Proc. of Amer. Acad, of Arts and Sc, VII, 1868, 368.

27. Groom, P. Contributions to the Knowledge of Monocotyledonous
Saprophytes. Journ. Linn. Soc. Bot., Vol. 31, 1894, 149.

28. Groom, P. On Thismia Aseroe (Beccari) and its Mycorhiza. Ann.
of Bot., IX, 1895.

29. Gruenberg, B. C. Some Aspects of the Mycorhiza Problem. Bull.
Torr. Bot. Club. 36, 1909.

30. Holm, Theo. Pyrola aphylla, A Morphological Study. Bot. Gaz.,

25.

31. Hooker, J. D. Flora of British India, 1875.

32. Irmisch, Th. Bemerkungen uber einige Pflanzen der deutschen
Flora Flora, 1855, 628.

33. Irmisch, Th. Einige Bemerkungen uber die einheimischen Pyrola
Arten. Bot. Zeit., 1856, 585.

34. Irmisch, Th. Kurze Mittheilung uber einige Pyrolaceen. Flora,

1859. 497-

35. Johow, Fr. Die chlorophyllfreien Humusbewohner West Indiens.

Jahrb. f. wiss. Bot., 16, 1885, 415.

36. Johow, Fr. Die chlorophyllfreien Humuspflanzen. Jahrb. f. wiss.
Bot., 20, 1889, 475.

37. Jussieu, A. L. Genera Plantarum, 1774.

38. Kamienski, Fr. Die Vegetationsorgane der Monotropa Hypopitys L.
Bot. Zeit., 39, 1881, 457.

39. Kamienski, Fr. Les Organes Vegetatifs die Monotropa Hypopitys L.
Mem. Soc. Nat. des Sciences Nat. et Math, de Cherbourg, XXIV, 1882.

40. Kerner, A. Natural History of Plants, II, 1895.

41. Klotsch. Der Grundlichste Kenne der Ericaceae und deren Ver-
wandten. Linnaea, XXIV, 1851, 5.

42. Knuth, P. Handbook of Flower Pollination, 1909.

43. Koch, L. Die Entwicklung des Samens von Monotropa Hypopitys L.
Jahrb. f. wiss. Bot., 13, 1882, 202.

44. Kramar, U. Studie iiber die Mikorrhiza von Pyrola rotundifolia.
Bull, intern. Acad. Sci., Prague, VI, 1901, 9-15.

108 Henderson — Comparative Study of Pyrolaceae and

45. Lidforss, B. Zur Physiologie und Biologie der wintergriinen Flora.
Bot. Centralblatt, XVII.

46. Lindley, J. Introduction to the Natural System of Botany, 1831.

47. Linnaeus, C. Species Plantarum, I, 2nd Ed., 1762.

48. MacDougal, D. T. Symbiotic Saprophytism. Annals of Bot., XIII,

1899. 31-38.

49. MacDougal, D. T. Symbiosis and Saprophytism. Bull. Tor. Bot.
Club. 26, 1899, 511.

50. MacDougal, D. T., and Lloyd, F. E. The Roots and Mycorhizas
of Some of the Monotropaceae. Bull. N. Y. Bot. Gard., I, 1900, 419.

51. MacDougal, D. T. Induced and Occasional Parasitism. Bull. Torr.
Bot. Club, 38.

52. McLean, R. C. The Utilization of Herbarium Material. New
Phytologist, 19 16.

53. Malme, G. O. Nagra bildningsafvikelser i blomman hos Pyrola
uniflora. Sv. Bot. Tidsskr., I, 9, 1907, 270-275.

54. Meehan, Th. The Evolution of Parasitic Plants. Bull. Torr. Bot.
Club, XVIII, 1891, 210-212.

55. Meehan, Th. On Some Morphological Distinctions in the Genera
of Ericaceae. Proc. Acad. Nat. Sc, Philadelphia, III, 1892, 372-374-

56. Niedenzu, F. Ueber den anatomischen Bau der Laubblatter der
Arbutoideae und Vaccinioideae. Bot. Jahrb., II, 1890, 135.

57. Nuttall, Thomas. The Genera of North American Plants, 1818.

58. Oliver, F. W. On Sarcodes sanguinea Torr. Ann. of Bot., 1890,
304-326.

59. Peklo, J. Die epiphytischen Mykorrhizen nach neuen Untersuch-
ungen. I Monotropa Hypopitys L. Bull. int. de l'Acad. de sciences de
Boheme, 1908.

60. Petersen, H. E. Arctic Flowering Ericaceae: The biological anatomy
of the leaves and of the stems. Meddelelser om Gronland, XXXVI, 1908.

61. Pursh. Flora Am., Sept., I, 1814, 300.

62. Queva, C. Le Monotropa Hypopitys L. Anatomie et Biologie.
Mem. Soc. d. Hist. nat. d'Autun, XXVII, 1909.

63. Rayner, M. C. Obligate Symbiosis in Calluna vulgaris. Ann. Bot.,
XXIX, 1915, 97.

64. Rayner, M. C. Recent Developments in the Study of Endotrophic
Mycorhiza. The New Phytologist, 15, 1916.

65. Rommel, W. Anat. Untersuchungen iiber die Gruppen der Piroleae
und Clethraceae. Heidelberg, 1898.

66. Roper, J. Normales und Abnormes. Bot. Zeit., 1852, 460.

67. Rylands. On the nature of the byssoid substance found investing
the roots of Monotropa Hypopitys. Phytologist, I, 1842, 341.

68. Schenk, A. Handbuch der Botanik, I, 1881. (Drude on Monotropa
Hypopitys, p. 604.)

69. Shibata, K. Die Doppelbefruchtung bei Monotropa uniflora L.
Flora, 1902, Bd. 90, 61.

70. Shibata, K. Experimented studien iiber die Entwicklung des Endo-
sperms bei Monotropa. Biologisches Centralblatt, 22, 1902.

Monotropaceae with Reference to Ericaceae 109

71. Simon, F. Beitrage zur vergleichenden Anatomie der Epacridaceae
und Ericaceae. Bot. Jahrb., 13, 1891, 15.

72. Small, J. K. Flora of North America, Vol. 29, Pt. 1, 1914.

73. Solereder, H. Systematic Anatomy of the Dicotyledons. Rev.
Ed., 1908.

74. Solms-Laubach, H. Ueber den Bau und die Entwicklung der Er-
nahrungsorgane parasitischer Phanerogamen. Jahrb. f. wiss. Bot., I, VI.

75. Stahl. Der Sinn der Mycorhizenbildung. Jahrb. f. wiss. Bot.,
XXXIV, 1900, 539.

76. Strasburger, E., and Hillhouse, W. Practical Botany, 1897,

330.

77. Ternetz, Charlotte. Ueber die Assimilation des Atmospharisches
Stickstoffes durch Pilze. Jahrb. f. wiss. Bot., 44, 1907, 353~4o8.

78. Torrey, J., and Gray, A. Flora of North America, 1838-40.

79. Torrey, J. Ann. Lye. N. Y., VII, 55 (1864).

80. Velenovsky, J. Ueber die Biologie und Morphologie der Gattung
Moneses. Rozpravy ceske Akad Rocnik I, Trida II, Cislo 39, 801, ff. 1892.

81. Velenovsky, J. Ueber die Keimpflanzen der Pirolaceen Ber. der
bohm. Kaiser Franz. Josefs-Akad. Jahrg., XIV, Kl. No. 35, 1905.

82. Vesque, J. Anatomie Comparee de L' Ecorce. Ann. d. Sciences
Nat., Ser. 6, Vol. 2, 1875.

83 Wakker. Untersuchungen iiber den Einfluss parasitischer Pilze auf.
ihre Nahrpflanzen. Jahrb. f. wiss. Bot., Bd. 24, 1892.

84. Warming, E. The Structure and Biology of Arctic Flowering Plants.
Ericineae. Medelelser om Gronland, XXXVI, 1908, 1-71.

85. Warming, E. On Skudbygnung Oberwintring og Foryngelse Natur-
historik Forenings Festskrift Kopenhagen, 1884, 86.

86. Warnstorf, C. Bot. Beobachtungen aus der Provinze Brandenburg
im Jahre 1894. Verh. Brand, XXXVII, 1895, 34-61.

87. Warnstorf, C. Bliithenbiologische Beobachtungen bei Neu Ruppen
im Jahre 1896. Zeitschr. naturw. Ver. Harz, XV, 1896, 9-20.

88. Breitfeld, A. Der Anatomische Ban der Laubblatter der Arbu-
toideae und Vaccinioideae. Bot. Jahrb., 9, 1888.

Bat. Contrib. Univ. Pain.

Vol. V, Plate I.

Russell on Sarracenia.

Bot. Con I rib. Univ. Pain.

Vol. V, Plate II

6

Russell on Sarracenia.

Bot. Contrib. Univ. Penn.

Vol. V. Plate III

T

^

i&

rac

*■

j"^""-

1JS

*» ..

«*« ' w

'xT

8

> ■

13

.j>

%.+ -%

iJfvV

*«e*-

^., ~

/J < -
h - J »

10

"., <r^r*>

I

18 ^ [i vj^ Lm

cs-^V

X

11

*/% " ifC Of r "1 / I V

^* ir

9

Russell on Sarracenia.

Sot. Contrib. Univ. Penn.

Vol. V, Plate IV.

* ■- A * <

I] *■■-■ : m

» *

, m

ii*«

, t ■" f

4* "!
•;.t.N i

16

}** ft ^

*/ f?

I-'-' ; a If

■mm ,t! '•■!•

fete

• ' r ;•'•- ■■■''

p.

■•a

Ippp ^

17

**• i

10

Mm®

m§mmm

t#* \ *' X

21

wmmm

20

Russell on Sarracenia.

Bot. Con t rib. Univ. Penn.

Vol. Y. Plate V.

yf\\ ■ •

24

23

» ! w" //

Russell on Sarracenia.

Vol. V 1920 No. 2

CONTRIBUTIONS

FROM THE

Botanical Laboratory

OF THE

University of Pennsylvania

University of Pennsylvania

Philadelphia

1920

CONTENTS OF VOLUME V, NO. 2.

Page
A Morphological and Gytological Study of Reproduction in
the Genus Acer. By William Randolph Taylor, B.S.,
M.S. (With plates vi, vii, viii, ix, x, xi,) 112

The Morphological Continuity of Scrophulariaceae and Oro-
banchaceae. By Irwin Boeshore, B.S. (With plates xii,
xiii, xiv, xv, xvi) 139

Vol. V 1920 No. 2

CONTRIBUTIONS

FROM THE

Botanical Laboratory

OF THE

University of Pennsylvania

University of Pennsylvania

Philadelphia

1920

A MORPHOLOGICAL AND CYTOLOGICAL STUDY
OF REPRODUCTION IN THE GENUS ACER

By William Randolph Taylor, B.S., M.S.
With Plates VI-XI.

CONTENTS

Introduction 112

Historical Review 112

Material and Methods 113

Floral Development 115

Pollen Development in Acer negundo 117

Reduction and Somatic Divisions in Other Species 122

Ovule Development 126

Fertilization and Embryogeny 129

Seedling Anatomy 131

Summary 133

Bibliography 134

Explanation of Plates 134

H2 Taylor — A Morphological and Cytological

LIHR-1
BOTANICAL

Introduction qakuB

At the inception of this work the writer planned to review the
genus Acer from the standpoint of the evolutionary tendencies
shown by the comparative morphology of the floral parts, as well
as the relative degrees of dicecism and flowering periods of the
different species. A preliminary survey of the available sources
of material and information showed that it would be imprac-
ticable to secure sufficient data to insure a comprehensive view
of all of the sub-genera. This was especially true with respect
to living material. On the other hand, the literature regarding
the development of the reproductive elements was found to be
very meagre, and in some particulars, contradictory. Since as
many species as could be handled in a cytological study were
within reach, an attempt to clear up the points in dispute and
to extend our knowledge of the genus to unstudied species led
to the abandonment of the original problem.

Historical

The first paper to appear dealing with the cytology of Acer
was by Ira D. Cardiff in 1906 (2). This was primarily a study
of synapsis and the organization of the heterotypic chromosomes,
presenting the viewpoint of parallel pre-synaptic threads and
paired chromomeres. He worked with Acer platanoides and sev-
eral other plants of widely differing groups, describing an essen-
tial similarity of behavior throughout. The reduced chromo-
some number for the maple considered is given as eleven.

This work was followed by a study of Acer negundo by Chester
A. Darling in 1909 (4), an effort to determine the basis for the
dioecious condition in this form. The whole history of the
maturation divisions as described by this author is at variance
with the usual process of pollen development. The chromatin
is said to leave the nucleolus during the presynaptic phases in
the form of round bodies which build up the spireme on the
linin threads. Later the spireme breaks up to form eight
bivalent chromosome segments, while from the nucleolus bud
off directly five more bivalents. Upon the breakdown of the
nuclear membrane what is left of the nucleolus is said to divide
up into a few granules which pass out from the spindle region.

05

CD

Study of Reproduction in the Genus Acer 1 13

The thirteen chromosomes thus formed divide equally and there
is found no basis in an unequal division for a differentiation
into male or female determining individuals. But Darling con-
siders that in the reorganization of the nuclei in the tetrad two of
the nuclei contain each a chromatin body lacking in the other
two, and that this feature may be related to sex determination.
The body in question seems to disappear soon after, during the
passage into the resting state.

The unusual conditions represented by this author as present
in Acer negundo were denied in a subsequent work by D. M.
Mottier on the same species, published in 1914 (13). Here a
very clear and convincing account is given of the conditions
present and a comparison made with the somewhat related
Staphylea trifolia L. He also, besides giving a more normal
history as that typical of the maple, disputes the correctness of
the count, giving twelve as the probable number in one place,
and in another suggesting either twelve or fourteen. Mottier
incidentally gives the chromosome count in Acer rubrum as
thirty-six in the reduced condition. In an earlier paper (12)
he describes the mature pollen of this species.

With regard to the ovule development, the only paper avail-
able is that of Mottier on Acer rubrum, 1893 (12). He reports
that late in March the megaspore mother-cell is evident as a
sub-epidermal cell, which becomes more deeply placed by later
divisions of the nucellar epidermis. Its first appearance was not
determined. The first division in the mother-cell is followed by
wall formation, and then the more micropylar of these again
divides with a resulting wall. Both these cells degenerate.
The more chalazal of the original pair divides later than the other,
but no wall is laid down, and the embryosac by two additional
successive divisions reaches the eight-nucleate stage. The three
antipodals early disappear after maturity, and the polar nuclei
fuse preliminary to fertilization.

Material and Methods

The writer was fortunate in being able to obtain material of
a considerable number of species, both native and exotic in ori-
gin. The greater part of the exotic material could not be worked
up with thoroughness, and so will not be considered in this paper.
Buds of Acer platanoides L. and Acer rubrum L. were made avail-

H4 Taylor — A Morphological and Cytological

able early in the spring of 191 8 by forcing in the greenhouses.
These were again collected at the normal flowering season, with,
in addition, Acer psendo-platanus L. Little was done with the
less common introduced species until the next season, when a
considerable series was obtained, as well as Acer negundo L.
buds, and the root tips of some species for vegetative mitoses.
In all, about two hundred fixations were made. The material
was found to be extremely hard to fix properly, a tendency being
evident for the chromosomes to clump at metaphase, which
caused many fixations to be discarded. The solution finally
adopted was a Strong Flemming containing less than one per
cent, of Urea and Maltose. The bud scales were first dissected
away and the flower mass plunged beneath the surface of alcohol
for a moment to drive the air from the interstices, from which
it was removed to the fixing solution. The large inflorescences
were separated into several parts to facilitate penetration.
Some species have resinous bud scales or tufts of hairs between
the flowers which render it almost impossible to get satisfactory
results. Later stages in the development of the ovules neces-
sitated the cutting away of the ovarian walls. Except in the
case of Acer pseudo-platanus, the root-tips used were obtained
from greenhouse cultures. They were fixed in Weak Flemming.
For pollen and vegetative cell studies sections were cut 4-5
microns thick, and for embryosac and embryo studies approp-
riate to the stages represented. The stain used throughout
the cytological part of the work was Heidenhain's Haematoxylin.
The writer is much indebted to the many friends who have
by contributions of living material or otherwise, assisted in this
work. To Dr. John M. Macfarlane acknowledgement is made
of the original suggestion and for the complete facilities
under which the work was conducted. For a large part of
the fresh material of Acer rubrum L., as well as other items, he
must thank Dr. Alice M. Russell. Through the kindness of
Mr. Roberts LeBoutillier the writer was able to collect much
material from Japanese species growing on his estate. Prof.
C. S. Sargent very generously sent a considerable number of
species from the Arnold Arboretum, but as the flowers in this
material were largely open, they will not enter into the present
paper to any great extent.

Study of Reproduction in the Genus Acer 115

Floral Development

The maples fall into two classes with respect to their time
of flowering, the one blooming before the development of the
leaves, the other after they have begun to expand. This fact
has been recognized in the subdivision of the genus (14). In
the region of Philadelphia Acer saccharinum and Acer rubrum
bloom within about ten days of each other, in late February or
early March. Then there is a period of ten days or more before
any other forms appear. These first two both flower before the
leaves appear, and so also does Acer negundo, which ushers in
the most active blooming period. A few of the exotic species
only, seem with us to delay beyond the latter part of April.

The early blooming of the three forms named, correlated as
it is with a reduced perianth and marked dicecism, would lead
one to suspect a possible difference from other species in the
period of pollen maturation and other reproductive phenomena.
Only one of these, however, shows such a condition. Acer sac-
charinum, which flowers earliest in the spring, matures its pollen
in the autumn. The writer followed the development of the
flower buds closely during September and October, 1919, and
found that development was very gradual till the time for the
reduction divisions came near. Then a rapid growth in size
accompanied the divisions. This phase came about October
20-24. The period of the formation of the tube and generative
nuclei appears to be in late winter, generally during early Feb-
ruary (Fig. 27). Flowering occurs from the latter part of Janu-
ary, by the earliest records available, to the middle of March
in exceptionally late seasons and sporadic cases. The great
majority of the trees flower together during the first thoroughly
warm spell of the year.

This species is followed by Acer rubrum, generally during the
early part of March, though, exceptionally, earlier. Here pollen
maturation does not take place in the fall, but during the open-
ing of the flower buds. The reduction divisions take place
during the swelling of the bud, and the tetrad stage is reached
as the anthers appear between the scales at the tip. The for-
mation of the tube and generative nuclei occurs just before the
rapid elongation of the filaments which accompanies anthesis.
The sperm nuclei were not seen: they are formed probably
subsequent to the shedding of the pollen.

n6 Taylor — A Morphological and Cytological

Acer negundo follows a week or ten days later than Acer rubrum,
and maturation of the pollen also occurs during the opening of
the buds. In these three forms, with relatively condensed in-
floresences it might be expected that in any given bud the stages
would be almost simultaneous, but a great variation, even in
the lobes of a single anther, was the regular condition. The
same holds true for Acer saccharum, which blooms a little later
than Acer negundo.

It is, however, in the forms with a more complex inflorescence
that this variation in the time of development reaches the great-
est degree. Acer platanoides, the next of the common forms
to open, will show any condition from mature archesporium
to young pollen grains in the same flower cluster. This renders
the accumulation of a considerable quantity of material of any
one stage rather a prolonged task. Maturation occurs during
the swelling of the bud, and by the time the scales have opened,
most of the anthers will contain pollen grains. The time for the
actual flowering of this species is generally the second week in
April. Acer pseudo-platanus has the largest flower mass of any
of the commonly grown species, and here development proceeds
from the base of the raceme toward the tip, with great secondary
differences in the stage of development of the flowers on the
secondary racemes. The reduction divisions take place largely
while the racemes are from 10-15 mm. in length. It flowers
soon after Acer platanoides. Data for the Japanese and the
more rarely cultivated European species are incomplete, but
they mostly flower, in this region, during April or early May.

It proved much more difficult to get information with regard
to the time at which the reduction divisions take place in the
megaspore mother cell. In general it was found that in a flower
with young microspore tetrads, the megaspore first prophases
were taking place, while the later phases followed the separation
of the tetrad.

Abnormalities in floral structure were not especially sought,
but some few appeared. With regard to the constancy of the
dioecious condition in Acer saccharinum and Acer rubrum it
should be noted that trees of both species bearing alike male
and female flowers were found. The flowers occurred in differ-
ent infloresences on the same twig in Acer saccharinum. Pre-
dominately male clusters of Acer platanoides often bore a few

Study of Reproduction in the Genus Acer 117

flowers, opening later than the others, in which the ovaries
were normal in appearance.

Details of abnormalities in the structure of individual flowers
were not noted, except the occurrence of tri-carpellary ovaries.
These were found in Acer rubrum, Acer saccharinum, Acer sacc-
harum, Acer platanoides, and Acer pseudo-platanus, the forms
in which quantities of female material underwent close exam-
ination.

Pollen Development in Acer Negundo

The early stages in the development of the anther and the
origin of the tapetum and archesporium were not traced. There
was some variation between the different species in the number
of layers forming the wall. An epidermal layer, and an outer-
most endothecial layer of larger cells were always present. In
addition inner wall layers were found, varying in number from one
which disorganized at maturity in Acer spicatum Lam. to three,
of which one persisted in the mature anther, in Acer pseudo-
platanus. The tapetum was always strikingly developed. It
was found to show a multinucleate condition of the cells in all
of the forms studied, and the nuclear divisions giving rise to this
state occurred both by a somewhat abnormal looking mitosis
and by amitotic fission. The latter was by far the commonest,
and the daughter nuclei often remained attached to each other,
forming a cluster near the center of the cell. Both began about
the time of synapsis, and in the mitotic type the number of
chromosomes appeared in excess of the normal sporophytic
number (Figs. 31-33).

The material at hand has permitted a fairly complete study
of five species with regard to the details of the reduction divi-
sions in pollen formation. These are Acer negundo, Acer rubrum,
Acer platanoides, Acer pseudo-platanus and Acer saccharum, in
addition to which fragmentary material of several other forms
was available. Of these, by far the best quality of material
came from the first named. The low number of chromosomes
in this species makes it a very satisfactory one for study, and
since the writer found some differences of detail from the descrip-
tion given by Mottier (13) a synopsis of the stages in Acer ne-
gundo will precede a comparison with the other forms studied.

n8 Taylor — A Morphological and Cytological

With the growth of the archesporial cell after the last vege-
tative division, the chromatic material becomes largely con-
fined to the nucleolus and only a delicate peripheral linin net-
work marked by a few chromatic granules can be distinguished
(Fig. i). With the approaching heterotypic prophase the num-
ber and size of these granules increase, but there is no sign of
the paired condition of threads and granules described by Car-
diff (2). As the leptonema network becomes strongly defined
the irregularity decreases somewhat, and it passes into synapsis
in the form of a net, not as a continuous spireme thread (Fig. 3).
The actuality of synapsis as a condition in the living cell has
been much questioned, especially by the animal cytologists (11),
but it is a constant feature in the present material (Fig. 4).
One thing seems sure, however: the more evidence there is of
poor fixation, the more the synaptic knot appears as a structure-
less black mass, and the less as an aggregation of threads. It is
quite conceivable that still further refinement of technic would
eliminate it, with consequent abandoning of its emphasis as an
explanation of various phenomena of inheritance, but the maples
are so resistant in the walls of their rather small anthers to the
penetration of fixing fluids that they would form poor material
on which to base a critical series of experiments.

From synapsis the spireme emerges in the form of loops (Figs.
5, 6), which extend to the periphery and result in a very distinct
hollow spireme (Fig. 7). The figures in this paper differ from
those given by Mottier in the greater delicacy and complexity
of the thread system, especially at the hollow spireme stage.
It is probable that this is due to a difference in fixation, and the
writer feels that the condition here figured is the more represen-
tative. There is good evidence of anastomosis between the
strands at this stage, and the longitudinal split present in some
plants could not be distinguished.

As shortening of the system during the approach of strep-
sinema proceeds, it becomes evident that the thick bands that
are to form the chromosome pairs are formed by the lateral
approximation of threads (Figs. 8, 9, 10). This opens up the
controversial question of the nature of the spireme, which has
recently been admirably presented by E. Digby (6). In a
general presentation such as this it seems well to withhold a
discussion of the matter, acknowledging, however, its important
bearing on current theories.

Study of Reproduction in the Genus Acer 119

There is no spiral twisting of the paired threads such as is
so characteristic of the stage in the lily (Fig. 11). Instead, they
break up into a number of segments equal to the reduced chro-
mosome count. These may appear to collect somewhat to one
side of the nucleus in a second contraction, but the writer is rather
inclined to believe that this appearance is in the nature of an
artifact. The halves of each pair now separate more or less,
forming very conspicuous rings (Figs. 12, 13), which by separa-
tion at the ends and contraction give rise to the chromosome
pairs, seeming to pass through a twisted stage during the short-
ening. Finally all elements contract to very short rods closely
associated in pairs lying around the periphery of the nucleus
(Fig. 14). The number of these is readily counted when few,
as in this case, but in those forms with many chromosomes, as
described later, the task becomes much more difficult.

The nucleolus has from the earliest beginnings of the division
kept its large size and quality of strongly retaining the stain.
Although for the most part apparently connected by strands
with the spireme, the writer would not consider that it con-
tributed chromatin material by bodily transfer, as has been
suggested (4), but rather through the intermediary of products
dissolved in the plasma and which are recombined in the spireme
into stainable chromatin. This is not an hypothesis readily
capable of demonstration, but it fits the observed conditions
better than the other.

With the approach of diakinesis as described, there appears
gradually a denser interior cytoplasmic zone surrounding the
nucleus, which eventually resolves itself into a complex system
of filaments that show signs of aggregation into sheaves by the
time the breakdown of the nuclear membrane occurs. At this
time the cytoplasm rapidly encroaches on the nuclear cavity,
strands pass into it, and the chromosomes become forced toward
the center (Fig. 14). The filaments then rapidly swing around
into a few sheaves showing as a multipolar spindle, and then
into two in the typical fashion. All this while the nucleolus
has been rapidly decreasing in size so by the time the bipolar
condition has been reached, the nucleolus has disappeared
(Figs. 17, 18). The chromosome pairs then become arranged
in the nuclear plate, the small rounded elements of each pair
being directed toward the poles, not lying side by side in the plane

120 Taylor — A Morphological and Cytological

of the equator (Fig. 15). This fact is of importance in connec-
tion with making the chromosome counts, for in a precise polar
view it eliminates the likelihood of confusing the count through
inclusion of both elements of a pair, since the member above
completely covers and hides from view its homolog below.

During anaphase the chromosomes pass to the poles as
small, more or less ovoid bodies (Fig. 16). They are figured
by Mottier as being elongate, and during the late anaphase as
showing the split that so frequently is to be observed in plant
chromosomes at this stage. In the writer's material, the better
fixed the material appeared, (as judged by lack of shrinkage and
especially by the wide and even distribution of the chromo-
somes in the plate), the less the chromosomes appeared elong-
ated in the anaphase condition. The only elongation that was
seen appeared to be due to a tendency of some of the chromo-
somes to stick together, probably a fault of fixation, and this
increased greatly in material in which clumping of the chromo-
somes was present. As for the split, that could not be distin-
guished by the writer in any of the species he examined. It is
to be noted that these bodies are extremely small: from one-
half to one micron only in diameter, and the writer has been
unable after long observation to find evidence of bipartition.

A count of the reduced number at metaphase or early ana-
phase is readily made in such a species as Acer negundo, which
fixes well and has a small number of chromosomes. This is for
the present case given by Darling (4) as thirteen, but in the later
paper by Mottier (13) this is questioned, twelve being stated
as the probable number, with, elsewhere, fourteen as an alter-
native. Why the intermediate count of thirteen should be dis-
carded when the investigator seems in doubt whether the num-
ber just above or just below that first given is more correct, the
present writer fails to see, since there is no especial reason to
expect an even number for the reduced count. The material
available for the present paper gave ample clear counts of thir-
teen to establish this number as the gametophyte chromosome
count (Figs. 51-56 inch).

When the chromosomes approach the poles, they tend to
spread out as flattened structures and to anastomose by projec-
tions from their edges (Fig. 20). This forms a bowl-shaped
structure open on the side facing the equator of the spindle

Study of Reproduction in the Genus Acer 121

(Fig. 19). The open space is filled with a clearer and less vacuo-
late plasma than the general neighboring region, clearly differ-
entiated from the spindle which it terminates. The nuclear
membrane first becomes evident over the external face of the
bowl, gradually extending around the sides and contracting the
open face, accompanied by the chromatin which becomes dis-
tributed over the entire periphery by the time the membrane
is complete. The reticulum then rapidly loses its definiteness,
but it does not seem to pass through a series of comminuting
stages that are in effect a reversal of the prophase stages. The
nucleolus in its earliest appearance is difficult to distinguish
from the large chromosome-derived bodies, but these stain
more lightly and become smaller the later the stage studied,
while the nucleolus gradually becoming larger, can soon be rec-
ognized (Fig. 21). This nucleolus in many cases seems to be the
only chromatic body at interkinesis, but deep staining shows
that there remain angular structures on the periphery about
equal in number to the chromosomes that went into the nucleus
(Fig. 22).

The second or homotypic prophase begins with the estab-
lishment of dark-staining angular bodies on the nuclear periph-
ery, seemingly from the centers just mentioned (Fig. 23). As
these increase in size the nucleolus shrinks, retaining its dark
staining capacity. They can often be counted, but the contour
of the nucleus and its small size reduces greatly the number of
favorable countable cases. After the break-down of the mem-
brane (Fig. 24) they pass from an irregular distribution into
the flat plate preparatory to division. The nucleolus, still
present, passes out from the forming spindle and takes up its
position where the apex of the heterotype spindle had formerly
been (Fig. 25). It retains its staining capacity up to this point,
and is a very conspicuous object. The equatorial part of the
heterotype spindle remains distinct through the homotype divi-
sion (Fig. 26). The nucleolus disappears suddenly, shrinking
and paling without evidence of fragmentation.

The axes of the two homotype spindles may lie in the same
plane, or in two planes at right angles to each other, or even in
any intermediate position, thus giving rise to tripartite or cruci-
ate tetrads, or intermediate forms.

After separation of the tetrad the elements enlarge consider-
ably, and each forms its individual wall that, as the pollen grain

122 Taylor — A Morphological and Cytological

nears maturity, shows three thin lines which after shedding
permit the emergence of the pollen tube. The mature pollen
grain has but two nuclei, the tube and generative nuclei, and
the latter remains undivided till after the pollen grain has been
shed. The appearance of the pollen grains of the various spe-
cies is similar, size being the only evident difference (Fig. 27).
The pollen grains of anthers in female flowers more often showed
abnormalities than in anthers of male flowers, and showed de-
generation as anthesis approached.

Pollen and Somatic Divisions in Other Species

The history of pollen development as given for Acer negundo
is typical in a general way of the genus. The differences ob-
served in other forms are for the most part those necessary to
effect the distribution of the larger number of chromosomes.

Amitosis was the most common form of nuclear division in
the tapetum cells, but mitotic division was observed in other
forms than Acer negundo, notably Acer pseudo-platanus. There
was no special differentiation of the cytoplasm near the spindle
in Acer negundo, but in Acer saccharum and Acer platanoides
a denser zone was present, slightly removed from the spindle
and surrounding it. In Acer rubrum such a dense area was pres-
ent in a pronounced form, frequently assuming a unilateral or
unipolar position (Figs. 29, 30). This appearance is hardly
likely to be an artifact, for the general fixation of the cytoplasm
and of the chromatin is quite good. In the forms with high
chromosome numbers there was naturally a greatly increased
complexity of the spireme at all stages, and especially was this
true of Acer rubrum. A complete discussion of certain peculi-
arities of this form are given later, but it is best mentioned
here that as strepsinema approaches there appears a much more
decided twiscing of the spireme than was found in Acer negundo
(Fig. 28). This was true to some extent of the other forms.
The extreme complexity of the spireme mass prevented a com-
plete study of Acer rubrum. Certain material of this species
showed a great deal of abnormality in the pollen grains, described
later, and in addition to these structural variations in some
cases delayed division of the nucleus into tube and generative
nuclei. Division of the generative nucleus within the grain was
not observed, either as a normal or as an abnormal occurrence.

Study of Reproduction in the Genus Acer 123

In Acer platanoides the reduced number of chromosomes is
given by Cardiff (2) as eleven. This the writer has been able
to confirm, although the amount of material available was not
great (Figs. 59-62). The seeds of this species were germinated
and the root tips fixed to verify this by making a 2x count.
Here a word must be recorded with regard to the conditions of
observation of the mitoses, especially the somatic ones. In all
of the maples studied the chromosomes in the vegetative parts
are very small indeed. The longest studied measures about
three microns in length and the smallest about one micron,
with in both cases a diameter of from two-thirds to one-third
micron. Even with good fixation, which with the roots was
consistently obtained, overlapping of the ends of these bodies,
and other confusing arrangements frequently occurred. This
made the interpretation of the complex plates a difficult matter,
and in the extreme cases of Acer rubrum and an abnormality
under Acer saccharinum, compels the count arrived at to be
considered merely as a very careful approximation. The method
pursued in making all the chromosome counts, both somatic
and gametophytic, was to draw the most perfect plates with the
camera lucida at a magnification of 3380 diameters, correct the
rough sketch by direct observation and recompare under the
camera before counting. This enabled the writer to be sure no
element had passed by unobserved. In the cases of small ix
plates, a few actual drawings were supported by many direct
observations, readilv made when a small number of elements
was concerned. With Acer platanoides, having the smallest
number of chromosomes yet observed in the genus, it was a
surprise to find that the somatic count obtained in the root tip
cells did not agree with the reduced count, for the number
twenty-six was persistently obtained (Figs. 57, 58). A very
few counts of twenty-five, twenty-seven and twenty-eight prob-
ably contain observational errors. The difference between the
observed number and that of twenty- two which was expected,
is too great to be an error, and indicates probably that there
exist varieties differing in cytological composition. As no sys-
tematic attempt was made to keep records of the exact trees
furnishing the material for each fixation, these features unfor-
tunately cannot be correlated with the horticultural forms and
varieties listed by Pax (14). A further feature of this kind
appeared and will be considered under Acer rubrum.

124 Taylor — A Morphological and Cytological

Acer pseudo-platanus was studied with respect to both the
vegetative and reduced counts, with the result that twenty-six
and fifty-two were found to be the ix and 2x numbers. The
root-tips of this species furnished especially favorable material
for counting, and no significant variations appeared (Figs.
63-66).

The Sugar Maple, Acer saccharum, showed few and large
chromosomes, thirteen being the observed ix number (Figs.
69-71).

Only one of the remarkable oriental maples, Acer carpini-
folium, will be noted here. This has, as the name indicates,
leaves resembling the Hornbeam, Carpinus. This condition
made it of interest to obtain a count, and the results indicate
that fifty-two is the probable somatic number in this case, as
in the last (Figs. 67, 68). The group was not shown to be
strikingly different cytologically, therefore, from the forms with
lobed leaves.

Because the rapidity with which the pollen maturation of
the Silver Maple occurred in the fall was unexpected, the writer
failed to obtain reduction division material that would furnish
the ix count. In a pollen grain a countable anaphase polar
view was however obtained, the two plates giving in one case a
clear twenty-six and in the other a more uncertain twenty-
seven as the chromosome number (Fig. 73). Germination of
the seeds gave material for the 2x count, which was determined
to be between fifty-one and fifty-five, in all probability fifty-
two as in the preceding cases (Figs. 72, 75). The plates and the
individual chromosomes of this species are the smallest observed,
if the number involved is considered. In two roots of this
form, Acer saccharinum, cases were found where the chromosome
number, while too great to be exactly determined, approxi-
mated a 4X condition at metaphase (Fig. 74). Most of the
plates in these roots were of the normal type, indicating that
this was an individual cell variation, probably by lack of separa-
tion of the daughter groups at the metaphase of a former divi-
sion. A condition believed to be similar, but not so clear, was
observed in the pedicels of Acer negundo.

The most perplexing situation presented, however, is that of
the Red Maple, Acer rnbrum. There is but one reference to the
cytology of this form, a sentence in Mottier's paper on Acer

Study of Reproduction in the Genus Acer 125

negundo. He there gives the reduced chromosome count for
Acer rubrum as thirty-six, the same as Staphylea. This marked
the form as strikingly different in nuclear composition from the
others then known, Acer negundo and Acer platanoides. It was
early examined by the writer for comparison with the others,
but the number of chromosomes was so large that with the poor
fh ation at first obtained, it seemed impossible to obtain a count.
With later material better success was attained, and the number
determined as being between sixty-eight and seventy-five (Figs.
30, 79-81). This is about twice the number given by Mottier.
It was not till material of the reduction divisions in the mega-
spore mother cell was studied that the original count of thirty-
six was seen (Fig. 34) by the writer. Here a few late prophases
gave results approximately verifying Mottier. The megaspore
metaphases were not in such a position that they could be
counted. This gives a very interesting situation, since the
pollen material first studied clearly represented a race with a
2x gametophytic count. As proof of two cytologically distinct
lines of Acer rubrum this evidence is far more decisive than that
for Acer platanoides. To add to the interest of the situation
some material belonging to two other batches was found, which,
though not as good as that from which the 2x gametophyte
count was obtained, showed clearly that the reduced number
of the chromosomes was here in the neighborhood of fifty (Figs.
82-84). If the original number of thirty-six is exactly correct,
which, although it seems probable, the writer would not under-
take to absolutely affirm with the few data at his disposal,
then seventy-two would be the expected reduced number in the
tetraploid form.

The pollen of the material above described probably repre-
sents such a race. If these hybridized, then the pollen of the
hybrid form would have a count of fifty-four according to ordi-
nary expectations, which was observed in one instance in the
two fixations just mentioned, although most of the counts run
from fifty to fifty-two. The seventy-two chromosome class of
pollen was found in pollen of male flowers, the apparently "hy-
brid" condition in the pollen of both male and female flowers.
Unfortunately, such a situation being totally unforseen, no
records are to be had that would locate the original trees. Meas-
urements were made to compare the size of the microspore

126 Taylor — A Morphological and Cytological

mother cells at the first metaphase. The longest and shortest
diameters of each grain were taken and averaged for the series.
The tetraploid material gave an average diameter for those
counted of 29 microns, the pollen from the "hybrid" male
flowers gave 23.5 microns, and the pollen from the female "hy-
brid" 24.8 microns. The difference between these last two is
hardly significant, being probably a shrinkage variation. The
volume of the tetraploid cells is therefore about 12,800 cubic
microns each, and that of the larger "hybrid" material about
8,000 cubic microns, which gives a ratio somewhat less than
that of the chromosome numbers, namely 4:2.5 as against 4:3.
Abnormalities in pollen formation have appeared in Acer
rubrum, which may throw light on the origin of these conditions.
In the material giving the intermediate count, one case of a tri-
polar spindle was found, with the chromosomes partlv oriented
with respect to one axis, partly with respect to the others. In
different material pollen grains were found of two sizes inter-
mixed, in diameter about 39 and 27 microns respectively. Fur-
thermore, conditions of lack of complete separation of the orig-
inal tetrad were found, both where, within the wall of the mother
cell, each of the four had formed a wall of its own and also where
they remained naked within the common investment. These
were in anthers where the majority of the pollen grains were
well advanced toward maturity. Some large grains were present
with abnormal numbers of nuclei. Germination of the seeds of
the Red Maple furnished material for a study of the somatic
divisions, but the extreme complexity of the figures made the
counting too much of a strain, and, since the accuracy was com-
parative at best, only a few cases were critically studied. The
number of chromosomes as counted seemed to be above ninety
(Figs. 76-78). Owing to the conditions mentioned it would
be unwise to try to interpret this in terms of the reduced number.

Ovule Development

The only maple which has received attention with respect to
the development of the ovule and embryosac is Acer rubrum,
upon which Mottier published a paper in 1893 (12). This paper
was written before some of the most important advances in
methods of technic, and it was thought advantageous to follow
the development again in the same species. The observations

Study of Reproduction in the Genus Acer 127

of the writer agree for the most part with those of the earlier
above-named investigator, but there are some points with regard
to which the writer would like to suggest a somewhat different
interpretation. To make these clear, a review of the situation
will be given.

The ultimate origin of the megaspore mother cell was not
determined. In material fixed about the end of October, 1918,
it was already differentiated as a deeply placed cell in the for-
ward end of the nucellus. The first ovular coat was just becom-
ing evident at that time (Fig. 85). Little advance in the growth
occurred till spring, when a rapid increase in size accompanied
the growth and maturation of the stamens, and the heterotypic
division in the megaspore mother cell occurred soon after the
microspore cells had reached the tetrad state. Both of the
ovular integuments were well developed by the time synapsis was
reached (Fig. 86). This applies only to the female flowers.
In the functionally male flowers the ovaries remain very small,
and the carpels seem to fail to close completely, leaving a rather
large opening at the stylar end. In the functionally female
flowers the carpels also remain sufficiently separate at the top
to leave a pore between the style bases, but this opening was
filled by the abundant stylar hairs, whose walls seem to become
partly gelatinized at this point.

Following the heterotypic division there is formed a wall,
separating the daughter nuclei into different cells (Fig. 87).
The writer was able to secure a very complete series of stages
illustrating the chromatin behavior during the heterotypic divi-
sion, which was fundamentally similar to the same series in the
pollen formation. In connection with the discussion of the
nuclear conditions in this species a megaspore prophase was used
to illustrate the condition where the reduced chromosome num-
ber was thirty-six, and the same figure, showing the thirty-six
paired segments nearing diakinesis, illustrates the difference in
size between the microspore and megaspore mother cell nuclei
(Fig. 34). So far there is no question as to the history of the
megaspore.

In Mottier's paper the description of the succeeding stages
is as follows: "The upper cell divides again in a similar manner,
so that there are three cells resulting from the mother cell (fig.
5). The lower one of these three now enlarges gradually ab-

128 Taylor — A Morphological and Cytological

sorbing the two upper; its large nucleus soon divides, and the
resulting nuclei move away from each other toward the opposite
ends of the cell (fig. 5). The further behavior of these nuclei
is similar to that which obtains in all known embryosacs of
angiosperms." From his figures it seems evident that the "upper
cell" means the one near the micropyle, and the "lower cell"
the one near the chalaza. His description would indicate that
the more micropylar cell goes through the homotypic division
before there is any division in the more chalazal cell, and that
following this division a wall is laid down and then the two cells
thus formed degenerate. The homotypic division in the chala-
zal cell of the pair then follows, and both of the nuclei formed
function in the maturation of the embryosac, each dividing
twice to give the eight elements of the mature sac. It appeared
to the writer an unusual circumstance that one cell of the orig-
inal diad should go through a complete homotypic division
only to degenerate, whereas in the other the nucleus only should
divide and both nuclei function. In view of the more recent
work in cytology it seems more to be expected that one of the
cells of the original diad should degenerate without dividing,
and that if the homotypic division in the other cell closed with
wall formation, the daughter cell adjoining the degenerating
cell should also break down. This would account for the two
degenerating cells reported by Mottier, and would indicate
that the nucleus in the remaining cell, morphologically equiva-
lent to one cell of a tetrad, would have to pass through three
successive free gametophytic divisions before the embryosac
reached maturity, which is rather the normal history in dicoty-
ledons. However, if Mottier actually saw the process of divi-
sion of the micropylar cell of the diad, this history could not
hold. But he does not figure any stage during the division,
and it is easily possible to misinterpret the condition if only the
final product of a row of three cells is available. The significant
difference between the two possibilities lies in that in the one
case only two gametophytic divisions occur, showing a tendency
toward the Lilium type where but one is present, while in the
second case the more normal number of three would be found.
The question as to whether a cell which is to undergo degenera-
tion is likely to divide earlier and in a more primitive manner
(i. e.} with wall formation) than its sister cell which is to

Study of Reproduction in the Genus Acer 129

function, also deserves consideration. Such material as the
writer has favors the interpretation here offered, but it is not
sufficiently conclusive to serve as a basis for discarding the state-
ment of Mottier, and thequestion must be left open for the present.

In any event, by the time the row of three cells is formed,
the ovule begins to bend downward, and the outer integument
on the lower side of the ovule begins to enlarge decidedly (Fig.
88). Bv the time the chalazal cell of the row of three has under-
gone the first division the ovule has increased greatly in size,
bent over nearly as far as it ever does, and it is evident that the
degeneration of the two cells at the micropylar end of the row
of three is far advanced. From the swelling of the outer integu-
ment the cells grow as hair-like processes (Fig. 89). The divi-
sion into the four nucleate stage shows these hair cells even more
elongate, and beginning to become crowded at the bottom of
the ovarian cavity (Fig. 90). They subsequently form a mass
of considerable size, and about the time of fertilization their
walls seem to be thickened in a gelatinous manner, and evidences
are present that the cavity of the ovary contains a partlv coag-
ulable liquid. The hairs may secrete a fluid which aids in the
passage of the pollen tube to the micropyle.

There is rarely any evidence of the degenerating cells bv the
time the eight nucleate stage of the embryosac has been reached.
Of these eight nuclei, three pass to the antipodal end and or-
ganize cells the membrane of which often is very indefinite.
Two form synergida? and one an egg at the micropylar end,
the latter protruding into the embryosac cavity above the
synergidse. Two nuclei remain, one at each end of the sac
(Fig. 36). These polar nuclei pass toward the center of the
cavity, where they come into contact (Fig. 37), and then pass
together to the egg, usually not fusing till they have taken up
a position close above it (Figs. 38, 39). Fusion is by simple
merging of the contents, no definite spireme being in evidence
at this time (8). This stage is generally not reached till after
the flowers have opened, and even till after pollination has taken
place. The antipodal cells very soon disappear, though thev
mav in occasional cases persist for some time (3).

Fertilization and Embryogeny
The fusion of the gametes in Acer, so far as observed, re-
sembles the conditions described bv Ishikawa for Oenothera (8).
The pollen tube in entering destroys usually only one of the

130 Taylor — A Morphological and Cytological

synergidae, the other persisting for a short time (Fig. 40). The
"filiform apparatus" was very slightly developed. As the flow-
ering of Acer saccharinum upon which the writer had counted
for the sequence of stages succeeding the entrance of the sperm
nuclei, was followed in 1919 by a severe frost and most of the
ovules rendered abnormal or infertile, a detailed discussion of
the stages in fertilization will be held for another paper. As
soon as the triple fusion is accomplished the endosperm nucleus
passes up to the middle of the embryosac (Fig. 40). Division
does not ordinarily take place till this position is reached (Fig.
42). The fertilized egg is even slower to divide, usually holding
back till the eight or even the sixteen nucleate stage of the endo-
sperm. The first wall is transverse (Fig. 41), and so is the
second, but the third is vertical or oblique in the terminal cell
(Fig. 43), and is followed by a similar wall in the middle cell of
the original row (Fig. 44). The basal cell seems to divide once
more (Figs. 45, 46), and in the later development of the embryo
may divide a few times to give the irregular group of cells which
form the suspensor of the embryo. The endosperm never be-
comes more than a sheath of protoplasm in which the free nuclei
divide, at first by mitosis, and later amitotically. They become
very numerous, reaching many hundreds. The cavity of the
embryosac increases greatly after fertilization, and the nucellus
stretches to accommodate it. This condition is especially
marked in Acer saccharinum and Acer platanoides. In the
latter the contents of the nucellar cells appear disintegrated,
and the walls themselves later in part break down, so that by
the time the embryo has reached the stage of the Red Maple
embryo shown in Fig. 47, the cavity of the embryosac will have
increased in volume a hundred times, and be surrounded by
only a very few layers of stretched nucellar cells, except just
above the micropyle, where the stretching and disintegration
does not occur. The embryo occupies only a very small part
of the pointed micropylar end of this cavity, and the rapid growth
of the sac and disintegration of the nucellus is therefore not due
to pressure from the embryo. The outer part of the second
integument becomes hardened, protecting the embryo, especially
in some of the exotic forms. The rapid growth of the embryo,
though delayed, causes the cotyledons in many species to become
crumpled in the seed. Evidence of the beginning of this con-

Study of Reproduction in the Genus Acer 131

dition appears long before the length of the embryo equals that
of the cavity, in Acer rubrum. The form of the embryo differs
somewhat among the various species even in the early stages
(Figs. 49, 50).

In most species the ovary did not grow much faster than
necessary to accommodate the swelling ovules, but in others,
Acer negundo especially, immediately after fertilization the
ovarian cavitv increased at a much greater rate than the ovules.
The growth of the alae is initiated by pollination, progressing
far before development of the embryo begins. Fertilization
seems to take place about forty to seventy-two hours after
pollination.

Abnormalities in embryosac development were many. It is
to be remembered that of the two ovules formed in each cavity
of the ovary only one comes to maturity, and that even in the
flowers that are developing normally half the embryosacs seen
will be in stages of abnormal development or degeneration.
The presence of sterile flowers increases the number of atypical
cases. The frost mentioned above seemed to destroy the egg
first, so that a several-nucleate endosperm with a shriveling egg,
as well as other degenerate conditions, was found in material
gathered on the days succeeding the cold spell.

In one ovule of Acer platanoides an embryosac with a double
embryo appeared. The second appeared as a smaller individual,
attached to the base of the primary one. It possibly resulted
from a bifurcation of the head end of the embryo while small,
after which one of the halves greatly out-grew the other (Figs.
91, 92).

Seedling Anatomy

A study of the anatomy of the developing Maple seedling
offers little of itself, but a comparison of the normal with a
tricotyledonous specimen which sprouted in one of the cul-
tures is of interest. The development in maples of specimens
with abnormal numbers of cotyledons and of later leaves in
whorls of three or even more is well known (15) as is also twin-
ning, etc. (5).

In Acer rubrum the cotyledons have in the contracted lower
part of the blade about six vascular strands (Fig. 93). These
unite in the petiolar part to two (Figs. 94, 95) which pass un-
fused into the hypocotyl (Figs. 95, 96). The plumule at first

132 Taylor — A Morphological and Cytological

shows only one strand in each leaf, but as they grow three be-
come evident which fuse in the petioles or, lower, in the hypo-
cotyl (Figs. 94, 95). The bundles which supply the plumule
become less and less prominent as one descends the axis, the
xylem disappearing and the phloem fusing laterally with the
adjacent bundles supplying the cotyledons (Figs. 96, 97), till
at last there is a simple axis with four collateral bundles (Fig.
98). These form the four poles of the tetrarch root by simple
rotation of the elements (Figs. 99, 100, 101).

In the same species the tricotyledonous specimen was pre-
pared in serial sections, and the vascular strands traced in the
same way. In the upper part of the cotyledons there were about
the same number of vascular strands as in the cotyledons of the
normal form. Toward the petiolar base in two of them the
strands partially united (Figs. 102, 103), to reappear again in
the hypocotyl as two distinct but closely placed bundles (Fig.
104). The third cotyledon possessed a small lobe near the
base (Fig. 102) which united with it before union with the axis
occurred. In this case the two bundles at the base remained
widely separated, a small patch of phloem tissue remaining
between them (Figs. 103, 104). There were three plumular
leaves, all prominently three-lobed. These had each three vas-
cular strands, which remained distinct into the hypocotyl, where
the xylem of the lateral members of each group disappeared,
and the phloem of these fused with the central member (Figs.
102, 103, 104). Instead of becoming united into four bundles
in the upper hypocotyl as in the normal specimens, the bundles
from cotyledons and plumule formed an irregular ring (Fig. 105).
A bend above the base of the hypocotyl made it impossible to
determine whether these fused to a regular number in the lower
hypocotyl, or, as seems likely from the irregular distribution of
the xylem (Fig. 106), merged in the transition zone at the base
of the hypocotyl. There the vascular system appeared as an
irregular ring, the xylem elements on the inner margin. In the
top of the root these passed out toward the periphery at four
points (Fig. 107) forming a tetrarch root that was somewhat
more irregular in form than the normal (Fig. 108).

The condition indicated seems to be that of a simple fission
of one of the two typical cotyledons carried to its greatest extent,
the bundles supplying each part being distinct far down the

Study of Reproduction in the Genus Acer 133

hypocotyl. The small lobe at the base of the third cotyledon
may or may not represent a partial development of the same
kind.

Summary

The Maples studied, with one exception, Acer saccharinum,
mature their pollen during the expansion of the buds in the
spring. Acer saccharinum matures its grains to the one nucle-
ate stage in the autumn, the division to form the tube and gen-
erative nuclei generally being delayed till shortly before flower-
ing. It is the first of the genus to bloom, being followed at inter-
vals by the other native forms, the exotic species closing the
series about the end of April.

Pollen development in Acer negundo proceeds much as de-
scribed by Mottier, and in the other Maples studied the process
is similar. The haploid chromosome number is thirteen. Hap-
loid counts were made in Acer platanoides showing eleven, in
Acer saccharum showing thirteen, in Acer pseudo-platanus show-
ing twenty-six, in Acer saccharinum showing twenty-six, and
in Acer rubrum showing approximately seventy-two, fifty-four
and (megaspore mother cell) thirty-six chromosomes.

Somatic counts were made in the root-tips of some species
with the following results: Acer platanoides, twenty-six; Acer
pseudo-platanus, fifty-two; Acer saccharinum, fifty-two, with also
isolated cells containing about twice that number; Acer car pinifo-
lium, fifty-two; and A cer rubrum, above ninety. The difference be-
tween the somatic and reduced counts in Acer platanoides mav be
due to differing strains with differing nuclear composition. The
conditions in Acer rubrum seem to indicate that a tetraploid
form exists, and the intermediate counts similarly seem to indi-
cate that hybrids with the normal diploid form occur.

Ovule development was followed in Acer rubrum and other
species, with in the species named a row of three potential
megaspores resulting from the divisions of the mother cell. The
chalazal one persists, the other two degenerate. The persistent
cell gives rise to the embryosac, which is normal with eight
nuclei. The polar nuclei fuse before fertilization and the endo-
sperm nucleus passes to the middle of the sac before dividing.
The embryosac enlarges greatly, especially in Acer platanoides,
and many free endosperm nuclei are produced, but no cell walls
are laid down. The growth of the embryo does not accompany

134 Taylor — A Morphological and Cytological

that of the ovular cavity, but is at first delayed, later progress-
ing more rapidly so as to fill the cavity often with accompany-
ing crumpling of the cotyledons.

Abnormalities of pollen formation, such as giant grains and
unseparated tetrads, were seen, and also abnormalities in the
formation of the embryosac. Tricarpellary ovaries appeared
in several forms, and a twin embryo was observed, with one
individual much larger than the other.

Bibliography

i. Allen, Charles E. The Basis of Sex Inheritance in Sphcerocarpos.
Proc. Amer. Phil. Soc, Vol. LVIII, p. 289. 1919.

2. Cardiff, Ira D. A Study of Synapsis and Reduction. Bull. Torrey

Bot. Club, Vol. 33, p. 271. 1906.

3. Coulter and Chamberlain. Morphology of Angiosperms. 1912.

4. Darling, Chester A. Sex in Dioecious Plants. Bull. Torrey Bot.

Club, Vol. 36, p. 177. 1909.

5. DeVries, Hugo. Mutation Theory. Vol. II. 191 1.

6. Digby, L. On the Archesporial and Meiotic Divisions of Osmunda.

Ann. Bot., Vol. XXXIII, p. 135. 1919.

7. Goebel. Outlines of Classification and Special Morphology. English

Translation. 1889.

8. Ishikawa, Rigakushi, M. Studies on the Embryosac and Fertilization

in Oenothera. Ann. Bot., Vol. XXXII, p. 279. 1918.

9. Kanda, M. Field and Laboratory Studies of Verbena. Bot. Gaz.,

Vol. LXIX, p. 54. 1920.

10. Mackenzie, Marion. Phytophenology in its Application to the Plants

of the Philadelphia Neighborhood. Trans. Bot. Soc. Penna., Vol. II,
p. 288.

11. McClung, C. E. Spermatocyte Divisions of the Acrididae. Kansas

University Quarterly, Vol. IX, p. 73. 1900.

12. Mottier, D. M. Development of the Embryosac of Acer rubrum. Bot.

Gaz., Vol. XVIII, p. 375- 1893.

13. Mottier, D. M. Mitoses in the Pollen Mother Cells of Acer negundo L.

and Staphylea trifoliata L. Ann. Bot., Vol. XXVIII, p. 115. 19 14.

14. Pax, F. Aceraceae in "Das Pflanzenreich", 8 Heft. (IV. 163).

15. Thisleton-Dver, Sir W. T. Morphological Notes. Ann. Bot., Vol.

XVI, p. 554. 1902.

Explanation of Plates

Plate VI — Stages in Reduction Divisions, Pollen Formation in Acer negundo.
All Figures Reproduced X2250.
Fig. I. Resting nucleus after last archesporial division.
Fig. 2. Nucleus showing approaching leptonema.

Study of Reproduction in the Genus Acer 135

Fig. 3. Leptonema entering synapsis as a network.

Fig. 4. Mid-synapsis.

Fig. 5. Looped pachynema threads leaving synapsis.

Fig. 6. Ditto.

Fig. 7. Hollow spireme.

Fig. 8. Contracting hollow spireme passing into strepsinema.

Fig. 9. Early strepsinema.

Fig. 10. Strepsinema, thread breaking up.

Fig. 11. Strepsinema, thread largely broken into chromosome segments.

Fig. 12. Ring-like chromosome segments.

Fig. 13. Ditto.

Fig. 14. Breakdown of the nuclear membrane, strands of cytoplasm
traversing the cavity.

Fig. 15. Polar view, heterotypic metaphase, showing thirteen chromo-
somes, haploid number.

Fig. 16. Side View, heterotypic anaphase.

Plate VII — Stages in Reduction Divisions, Pollen Formation in Acer negundo
(Figs. 17-26), Acer rubrum (Figs. 28-30), and Pollen Grain of Acer
saccharinum (Fig. 27). All except Fig. 27 reproduced X2250.

Fig. 17. Multipolar spindle, early with large nucleolus.

Fig. 18. Multipolar spindle, later, with smaller nucleolus (in center).

Fig. 19. Side view, telophase of heterotypic division.

Fig. 20. Polar view, telophase of heterotypic division, showing thirteen
chromosomes.

Fig. 21. Late telophase, showing formation of nucleolus, the large body
in center of nucleus.

Fig. 22. Nucleus during interkinesis.

Fig. 23. Prophase of homotypic division.

Fig. 24. Polar view, breakdown of nuclearmembrane, homotypic divi-
sion, nucleolus evident.

Fig. 25. Multipolar spindle, homotypic division, showing the remains of
the homotypic spindle, and the nucleoli near the periphery of the cell.

Fig. 26. Homotypic division, side and polar views, showing remains of
heterotypic spindle and thirteen chromosomes in the polar view of the
metaphase.

Fig. 27. Pollen grain, mature state, with tube and generative nuclei,
Acer saccharinum. X950.

Fig. 28. Strepsinema, Acer rubrum.

Fig. 29. Side view, heterotypic metaphase, showing chromosome pairs
and polar position of the dense mass of cytoplasm. Seventy-two
chromosome form.

Fig. 30. Polar view, heterotypic metaphase, showing lateral position
of the dense cytoplasmic mass. Shows count of seventy-two chromo-
somes.

Plate VIII — Tapetal Cells of Acer negundo (Figs. 31-33). Embryosac and
Embryo Formation in Acer rubrum (Figs. 34-41, 43-49). Embryosac
and Embryo of Acer saccharinum (Figs. 42, 50). Reproduced X200,
except as indicated.

136 Taylor — A Morphological and Cytological

Fig. 31. Tapetum cells, showing amitotic nuclear division. X900.

Fig. 32. Ditto. X900.

Fig- 33- Tapetum cells, showing mitotic nuclear division. X900.

Fig. 34. Megaspore mother cell, early diakinesis, showing about thirty-
six chromosome pairs. X2250.

Fig. 35. Row of three megaspores, micropylar pair degenerating, chalazal
undergoing nuclear division.

Fig. 36. Embryosac before migration of the polar nuclei to the center
of the embryosac.

Fig. 37. Meeting of the polar nuclei at the center.

Fig. 38. Fusion of the polar nuclei after reaching the egg. The nuclear
membranes between the nuclei have disappeared. X470.

Fig. 39. Mature embryosac ready for fertilization.

Fig. 40. Embryosac shortly after fertilization. One of the synergida>
has been destroyed, the other is disintegrating but the nucleus is evi-
dent. The dark mass introduced by the pollen tube is present,' and
the endosperm nucleus has begun to move toward the center of the em-
bryosac.

Fig. 41. Two cell stage of the embryo, eight nucleate endosperm.

Fig. 42. Fertilized embryosac before division of endosperm nucleus.

Fig. 43. Four-celled embryo.

Fig. 44. Five-celled embryo.

Fig. 45. Nine-celled embryo.

Fig. 46. Eight (?) celled embryo.

Fig. 47. Older embryo.

Fig. 48. Embryo showing beginning of the two cotyledons. X25.

Fig. 49. Older embryo of Acer rubrum. X25.

Fig. 50. Older embryo of Acer saccharinum. X25.

Plate IX — Chromosome Counts in Various Maples. All figures repro-
duced X2250.

Fig. 51. Upper and lower groups, heterotypic anaphase in Acer negundo.
Thirteen chromosomes in each case.

Fig. 52. Metaphase plate, homotypic division, Acer negundo, showing
thirteen chromosomes.

Fig. 53. Ditto, another case.

Fig. 54. Metaphase plate, heterotypic division, Acer negundo, showing
thirteen chromosomes.

Fig. 55. Ditto, another case.

Fig. 56. Ditto, another case.

Fig. 57. Metaphase plate, division in root-tip, Acer plalanoides , showing
twenty-six chromosomes.

Fig. 58. Ditto, another case.

Fig. 59. Metaphase plate, heterotypic division, Acer platanoides, show-
ing eleven chromosomes.

Fig. 60. Ditto, another case.

Fig. 61. Anaphase group, heterotypic mitosis, Acer platanoides, showing
eleven chromosomes.

Fig. 62. Ditto, the other group from the same cell.

Study of Reproduction in the Genus Acer 137

Fig. 63. Metaphase plate, division in root-tip cell, Acer pseudoplalanus,

showing fifty-two chromosomes.
Fig. 64. Ditto, another case.
Fig. 65. Metaphase plate, heterotypic division, Acer pseudo-plalanus,

showing twenty-six chromosomes.
Fig. 66. Ditto, another case.
Fig. 67. Metaphase plate, division in root-tip cell, Acer carpinifolium,

showing apparently fifty-five chromosomes.
Fig. 68. Ditto, another case, with the more frequently found number of

fifty-two chromosomes.
Fig. 69. Metaphase plate, heterotypic division, Acer saccharum, showing

thirteen chromosomes.
Fig. 70. Anaphase plate, heterotypic division, Acer saccharum, showing

twelve chromosomes.
Fig. 71. Ditto, other plate in small cell, showing thirteen chromosomes.
Fig. 72. Metaphase plate, division in root-tip cell, Acer saccharinum,

showing fifty-two chromosomes.
Fig. 73. Anaphase groups, division in pollen grain to form tube and

generative nuclei, Acer saccharinum, showing twenty-six and (?)

twenty-seven chromosomes.
Fig. 74. Metaphase plate, division in root-tip cell, Acer saccharinum,

showing ninety-one (or more) chromosomes, approaching twice the

normal number.
Fig. 75. Ditto, ordinary cell, showing fifty-two chromosomes.
Fig. 76. Metaphase plate, division in root-tip cell, Acer rubrum, showing

ninety (or more) chromosomes.
Fig. 77. Ditto, showing eighty-eight (or more) chromosomes.
Fig. 78. Ditto, showing ninety-four (or more) chromosomes.

Plate X — Chromosome Counts and Ovule Development in Acer rubrum,
Abnormal Embryo in Acer platanoides (Figs. 79_84 X2250, Figs.
85-90 XI35)-

Fig. 79. Metaphase plate, heterotypic division, showing seventy chro-
mosomes.

Fig. 80. Ditto, showing sixty-eight chromosomes.

Fig. 81. Ditto, showing sixty-seven chromosomes.

Fig. 82. Ditto, showing fifty-three chromosomes.

Fig. 83. Ditto, showing fifty-four chromosomes.

Fig. 84. Anaphase groups, heterotypic division, showing fifty-two and
forty-eight chromosomes.

Fig. 85. Ovule, resting condition in late autumn.

Fig. 86. Ditto, at time of synapsis.

Fig. 87. Ditto, after heterotypic division, showing two cells derived from
megaspore mother cell.

Fig. 88. Ditto, showing row of three potential megaspores.

Fig. 89. Ditto, showing micropylar two in row degenerating, chalazal
cell in the two nucleate stage of the embryo-sac.

Fig. 90. Ditto, embryosac cell in four nucleate stage.

138 Taylor — A Morphological and Cytological

Fig. 91. Twin Embryo, Acer platanoides, showing cells of the smaller

embryo. X200.
Fig. 92. Ditto, showing attachment region of the embryos. X200.

Plate XI — Dicotyledonous and Tricotyledonous Seedling Structure in Acer

rubrum. All Figures X30.
Fig. 93. Dicotyledonous seedling section through lower part of blade of

cotyledons.
Fig. 94. Ditto, section through upper part of cotyledonary petioles.
Fig. 95. Ditto, through junction of cotyledons with hypocotyl.
Fig. 96. Ditto, just below cotyledonary node.
Fig. 97. Ditto, a little lower than Fig. 96.
Fig. 98. Ditto, through middle region of hypocotyl.
Fig. 99. Ditto, at top of transition region near base of the hypocotyl,

showing bundle ring.
Fig. 100. Ditto, lower, showing passage of xylem toward the periphery

at four points, in the top of the root.
Fig. 101. Ditto, lower, in main root axis, showing tetrarch character.
Fig. 102. Tricotyledonous embryo, section through lower part of blade

of cotyledons.
Fig. 103. Ditto, through junction of cotyledons with hypocotyl.
Fig. 104. Ditto, just below cotyledonary node.
Fig. 105. Ditto, through middle region of hypocotyl.
Fig. 106. Ditto, through top of transition region near base of the hypo-
cotyl, showing bundle ring.
Fig. 107. Ditto, lower, showing passage of xylem toward the periphery at

four points, in the top of the root.
Fig. 108. Ditto, lower, in main root axis, showing irregular tetrarch

structure.

Note — All drawings of cytological details except stages in the develop-
ment of the embryosac were drawn at a magnification of 3,380 diameters.
Other details were drawn at lesser magnifications, and all figures except those
on Plate XI have been reduced in reproduction to two-thirds of the original
size. The figures of seedling structure were drawn at 75 diameters and re-
duced to 30 diameters in reproduction.

THE MORPHOLOGICAL CONTINUITY OF
SCROPHULARIACEAE AND OROBANCHACEAE

BY

IRWIN BOESHORE, B. S., Ph. D.
With Plates XII-XVI.

(Thesis presented to the Faculty of the Graduate School in

partial fulfilment of the requirements for the degree

of Doctor of Philosophy.)

CONTENTS

Introduction 139

Review of Previous Literature and Discussion 141

General Morphology of the Root 147

Histology of the Root 15°

Comparative Morphology and Physiology of Stem and Leaf 153

Comparative Study of the Inflorescence 160

Comparative Study of the Flower 161

The Calyx 161

The Stamens 164

The Pistil 167

The Ovary 167

Hairs of the Pistil 167

The Nectary 168

The Seeds 169

Selection of Hosts I7°

Summary I7I

Conclusions *74

Explanation of Plates *75

Literature cited 1 76

139

140 Boeshore — The Morphological Continuity of

Introduction

In the past, the genera of Orobanchaceae have been regarded
as of nearest affinities with the Gesneraceae; this wholly because
of the one-celled ovary. But such a proceeding only takes ac-
count of one morphological detail. Is there any other family
with which Orobanchaceae can be directly and continuously
connected in many morphological features? The group, being
wholly parasitic, must of necessity be derived as an offshoot from
some family wholly free or more or less parasitic in habit.

The only related family showing such is Scrophulariaceae.
While the majority of these in genera are green and independent
in their nutrition, others show all stages in degradation-transi-
tions to reduced green root parasites, thence to yellowish, or
red-yellow parasites, finally to degraded simplified parasites
that are wholly heterotrophic (Harveya, Hyobanche) . Thus, for
example, in the single genus Gerardia of the Eastern United
States some species are tall leafy plants, as in G. flava, and these
show slight root parasitism. Transitions can be traced from
this to G. pedicularia, thence to G. purpurea, to G. oligophylla,
and to G. aphylla, the last of which has scant twigs and small
leaves that may be almost absorbed. The flowers in the latter
are few and reduced in size, while parasitism is almost complete-
ly established in relation to other plants. From such forms
transition is easy and gradual to Striga, Harveya, and Hyobanche.
But between Harveya and G. aphylla on the one hand, or species
of Orobanche on the other, the macroscopic and microscopic
resemblances, as will afterward be shown in this paper, are grad-
ed and complete. Is then the fundamental resemblance to
Gesneraceae of a one-celled ovary as compared with Scrophular-
iaceae with its two-celled ovary fundamental and indicative of
true affinity? This question will be fully worked out in a later
part of the paper. But attention might now be drawn to .ome
comparative morphological facts that may shed light on the
present problem.

If we review the more pronounced saprophytic families of
flowering plants, like Burmanniaceae, Orchidaceae, Ericaceae,
(including Monotropaceae) and Gentianaceae, all show degra-

Scrophulariaceae and Orobanchaceae 141

dation-transitions from green strongly vegetative plants of in-
dependent nutrition, to others with feeble root saprophytism
and thence to highly degraded colorless and heterotrophic plants
like Thismia of Burmanniaceae, colorless genera like Epipogon,
Aphyllorchis, Neottia, and Corallorhiza in Orchidaceae, color-
less Monotropa and other genera of the Ericaceae, and the color-
less Leiphaimos in Gentianaceae. Now in all of these, distinct
and continuous transition is seen from five-celled, three-celled,
or two-celled states (Ericaceae, Burmanniaceae, Orchidaceae,
two-celled Gentianaceae) of the ovary to incompletely five-three
or two-celled, and ultimately a one-celled state with parietal
placentation. Such might, therefore, suggest that Orobanch-
aceae simply represents a greatly degraded offshoot series or
sub-family of parasitic habit that has gradually been derived
from Scrophulariaceae in which slow absorption of the ovarian
partitions has resulted in a one-celled state from a primitively
two-celled. If such be true, then varied morphological trans-
sitional characters should be traceable between Scrophular-
iaceae and Orobanchaceae, while correlated with this a graded
physiological parasitism and degradation should also be observ-
ed. It is the aim of the present thesis to demonstrate that such
is correct.

In undertaking this, the writer might first review the varied
opinions held by previous observers.

Review of Literature and Discussion

Wettstein (1, p. 48) separates the Scrophulariaceae from the
Orobanchaceae and Gesneraceae on account of the one-celled
ovary in the last two. He states that a great many Scrophu-
lariaceae have close affinities with other families. "Die Orobanch-
aceae und Gesneriaceae lassen sich durch den 1 facherigen Frkn.
mit parietaler Placentation .... von den Scrophular-
iaceae unterscheiden. " He then states that clear connections
are shown in the genera Harveya, Hyobanche, and Buchnera of
Scrophulariaceae with Orobanchaceae.

Beck (2, p. 128) makes this statement: "The Orobanchaceae,
which are frequently regarded as a parasitic side line of the Ges-
neraceae, are separated from the Gesneraceae through the per-
fect superior fruit, from the Cyrtandreae through the richly de-

142 Boeshore — The Morphological Continuity of

veloped endosperm and the undeveloped endosperm of the seed
. . . . from both especially through their parasitism.
The one-celled fruit separates them from the Scrophulariaceae
with which they have also much in common." (Present Au-
thor's Trans.)

Fritsch (3, p. 141) under the Gesneraceae treats the connec-
tions of Gesneraceae with related families, especially with the
Scrophulariaceae, Orobanchaceae and Bignoniaceae, and holds
that a sharp distinction can scarcely be drawn between these
families. He further says, "Hingegen stehen die Orobancha-
ceae den Gesneraceae so nahe, dass die Auffassung derselben als
einer parasitischen, laubblattlosen Unterabteilung der Gesner-
iaceae keinen grossen Fehler involvieren diirfte. Immerhin ist
die Placentation und der Bau des Frkn. iiberhaupt ein Unter-
scheidungsmal zwischen den Gesneriaceae, Orobanchaceae und
Scrophulariaceae. "

Baillon (4) makes no comparison of the Scrophulariaceae with
the Gesneraceae and Orobanchaceae. He does not consider the
Gesneraceae and Orobanchaceae as two separate families, but
regards the Orobanchaceae (of other authors) as a parasitic ser-
ies of the Gesneraceae.

LeMaout and Decaisne (5, p. 593) include the genus Hyo-
banche in the Orobanchaceae, though on account of its two-celled
ovary, most authors place it in the Scrophulariaceae. The fol-
lowing is stated: "Orobanchaceae approach the Scrophulari-
aceae in their regular corolla, didynamous stamens, capsular
fruit, and albuminous embryo; they differ in their leafless and
scaly stem and parietal placentation. This placentation, their
glandular disk, and the preceding characters ally them to Ges-
neraceae, from which they are separated by their scattered
scales, parasitism, hypogynous corolla, and basilar embryo."
Warming (6, p. 525-28) places Lathraea in the family of Scro-
phulariaceae. The plant is described as pale yellow, or reddish
(without chlorophyll) ; it is parasitic on the roots of the Hazel,
Beech, and other shrubs or trees, having an aerial stem, and an
underground perennial rhizome .covered with opposite, scale-
like, more or less fleshy leaves. It approaches Gesneraceae in
having a unilocular ovary with two parietal placentae. With
him Orobanchaceae finds no place as a family, the genus Oroban-
che being included in the Gesneraceae. "Orobanche (Broom-
rape) is allied to this order as a parasitic form."

Scrophulariaceae and Orobanchaceae 143

Perhaps the most extensive morphological and physiological
studies of parasitic Scrophulariaceae and Orobanchaceae have
been made by Heinricher (7, p. 390-451, 665-773) and Solms-
Laubach (8, p. 560-75) on the former, by Beck (9, p. 7-70) and
Koch (10,) on the latter. The work on such genera as
Euphrasia, Odontites, Pedicularis and not least Bartsia and
Tozzia, by Heinricher has a very definite bearing on the present
investigation. To quote from him on Bartsia and Tozzia: "In
der That sind es diese beiden Rhinanthaceen, welche uns die
Briicke von den halbparasitischen Rhinanthaceen zu der holo-
parasitischen Gattung Lathraea bauen. . . . Der Aufbau
des Lathraea- Rhizoms ahnelt sehr dem von Bartsia, die Unter-
schiede sind wesentlich dadurch bedingt, das erstere H0I0-,

letztere Hemiparasit ist Tozzia nimmt eine

ganz eigene Stellung in der Rhinanthaceen-Rheihe ein; sie ist
nicht Holoparasit und nicht Hemiparasit, sondern sie ist beides
in zeitlicher Folge. Und so wird sie eben zum biologischen
Bindeglied Zwischen den Halbschmarotzern und der Holopara-
sitischen Gattung Lathraea. "

In view of the above comparative estimates of different au-
thors, the question at issue resolves itself into one of three po-
sitions. (1) The Gesneraceae and Orobanchaceae are most
nearly related to each other in that they both possess a one-celled
ovary with deep to shallow parietal placentation. (2) Oro-
banchaceae and Scrophulariaceae are most nearly related to each
other in their root parasitism, their alternate or at times opposite
leaves, their progressive parasitic degradation, condensation of
axis, and eventually non-chlorophylloid aspect. They only
differ in the two-celled ovary, but Scrophulariaceae seem to be
united with Orobanchaceae by Christisonia neilgherrica, with
its two-celled, becoming above one-celled ovary, as well as by
Lathraea with its imperfectly two-celled ovary. (3) The Oro-
banchaceae stand by themselves as a family.

To take the last caption first, Orobanchaceae being wholly
parasitic and non-chlorophylloid, clearly suggests that physiolog-
ically and now morphologically it is a degenerate offshoot from
some family that tended gradually to show semi-parasitic habit.
Now of all the Bilabiatae, there is only one other family which
shows, like Orobanchaceae, parasitic habits, namely, Scrophul-
ariaceae. No Gesneraceae are parasitic, or even show a slight

144 Boeshore — The Morphological Continuity of

beginning of parasitism. Such being the case, the second be-
comes that which we are inclined to favor. But the two-celled
. ovary of Scrophulariaceae and the one-celled ovary of the Oro-
banchaceae has been the barrier to such a connection with mor-
phologists in the past.

As a preliminary, therefore, to subsequent studies and inves-
tigation it may be profitable to compare groups of parasitic and
saprophytic plants already known to us. To start with sapro-
phytic families first, it is well known as already shortly stated
above, that in the Burmanniaceae the green and least degraded
genera, such as Burmannia, have three-celled ovarv with cen-
tral placentation, while in Gymnosiphon and the most degraded
genera like Thismia and Arachnites the ovary has become one-
celled with deep to shallow parietal placentas.

In the large family of the Orchidaceae as now recognized by
systematists the most primitive subdivision Apostasieae, has
subregular flowers, three to two stamens, and a three-celled ov-
ary to the pistil. In the more evolved subdivision Seleniped-
ieae, the flowers are decidedly irregular, the stamens are two in
number, and the ovary is three-celled. Root saprophytism in
its commencing stages is amongst these not uncommon. In the
division Cypripedieae, the stamens are still two in number, but
the ovary usually has become one-celled by varying stages of re-
duction in ingrowth of the carpellary margins by graded steps
that can well be traced. Saprophytism is a frequent feature of
their roots. In the division Orchideae, the flowers are most
varied and highly specialized, the stamens are now reduced to
one functional, the ovary is one-celled with shallow parietal pla-
centas. In this division all stages of condensing degradation
and loss of chlorophyll can be observed, till such non-chloro-
phylloid and greatly degraded genera like Neottia and Corall-
orhiza are reached. Unquestionably here a continuous process
of condensing reduction and degradation in the above more or
less related genera is correlated with increasing saprophytism,
all of this being associated with a gradual transition from three-
celled ovary to one-celled ovary with deep placentas and thence
to one-celled ovary with shallow marginal placentas.

Again, the family Ericaceae when treated in the only appro-
priate morphological manner that explains the evolution of the
subdivisions satisfactorily, includes primitive shrubby plants

Scrophulariaceae and Orobanchaceae 145

with ample green leaves that have evolved along at least three
main lines. One line, as beautifully traced by Dr. Henderson
(11) in a recent paper, shows condensing and degrading simpli-
fication through saprophytism till low herbaceous shrubs with
few scattered leaves like Chimaphila are reached, while these
lead to the colorless greatly degraded genera of the Monotro-
paceae like Sarcodes and Monotropa. Still another line leads
to the parasitic Lennoaceae, the carpels of which have probably
increased in number by subdivision of a primitive five, that re-
main as many celled but the ovules have become reduced to two
or one in each cavity. A third line passes to highly evolved
types like the Azaleas, Kalmias, and Rhododendrons with ample
leafage, and roots that are nonsaprophytic or only slightly sa-
prophytic and ovary that is five-celled.

In the Gentianaceae, that frequently tend to be slightly or
markedly saprophytic, the ovary is still two-celled in the subdi-
vision Exacineae. In others like Lisiantheae the ovary is strict-
ly one-celled but has deep almost adjacent placental ridges. In
the Gentianeae the one-celled ovary has shallow ridges that
reach a climax of shallowness in the colorless degraded sapro-
phytes Voyria and Leiphaimos.

In the Convolvulaceae the ovary throughout is typically two-
celled and this is retained in species generally of the parasitic
genus Cuscuta, but a one-celled parietal condition is present in
Erycibe. Similarly in Rafnesiaceae, the ovary is one-celled.

From the above it is abundantly evident that alike in sapro-
phytic and parasitic families simplifying and degrading sapro-
phytism and parasitism nearly always are accompanied by a
transition from a several-celled central type of placentation to
a one-celled parietal type.

Accepting the foregoing as an indubitable fact, the question
may now be asked: Can close morphological affinity be traced
from Scrophulariaceae through types that become more or less
parasitic and eventually degraded colorless parasites, to forms
typical of the purely parasitic and non-chlorophylloid Orobanch-
aceae? The demonstration of this is the main thesis of the pre-
sent paper. As strongly favoring such a position, it may be
noted that in such a genus as Gerardia the most gradual transition
can be traced from tall leafy green species, like G.flava (Aureo-
laria villosa) to G. pedicularia, thence to G. purpurea, to G. oli-

146 Boeshorer—The Morphological Continuity of

gophylla with few pale green leaves and branches that mainly
carry on food elaboration, and finally to the most degraded mem-
ber, G aphylla, in which the leaves are small pale green scales,
still, however, provided with a few stomata, but which are in the
last stages of disappearance. It is of interest to note the close
resemblance in habit, structure, etc., between the two latter types
and such South African genera as Striga and Harveya, which
connect again with Hyobanche. All of these, but the last in par-
ticular, closely simulate members of the so-called Orobanchaceae.
Further, it is important to note that in the genus Christisonia
(Campbellia) , the ovary, while mainly one-celled with deep pla-
cental ledges, is in most of the nine species of Christisonia two-
celled in the lower part. And so Beck (12, p. 131) has well not-
ed, "Ausnahmsweise kammt bei letzterer ein unteren Teile 2
facheriger Frkn. vor; es besteht somit eine starke Annaherung
an die Gatt. Harveya." Regarding Christisonia neilgherrica
Worsdell (13, p. 131) says, "The ovary is bilocular in its lower
region and unilocular above; in the latter case the placentation
is parietal. In this plant the basal portions of the two bipart-
ite placentas very nearly meet in the centre. In the lower, bi-
locular part of the ovary, where the projections have become
united to form a dividing wall, the placentation is axile, two
placentas, bearing a large number of minute ovules, projecting
into each cavity."

If the above morphological and physiological lines of con-
tinuity express a correct interpretation of the lines of evolution
pursued, it should be possible to trace certain fairly continuous
morphological similarities as well as degradation-differences, pro-
ceeding in relation to the following parts.

First, since the phenomenon is fundamentally due to increas-
ing root parasitism, we should expect to find that the root and
gradually thereafter the vegetative stem system should show
condensation and swelling of the condensing axis, or, that this
axis should become rather starved and simplified.

Second, in such condensed axes an increasing preponderance
in relative width and importance of phloem over xylem should
occur until in the most condensed genera it would become a pre-
ponderant feature.

Third, in the above process and with the increasing parasitic
degradation the leaves would tend by degrees to become reduc-

Scrophulariaceae and Orobanchaceae 147

ed in size, and in nutritive capacity, till they would become re-
duced to relatively small colorless scales.

Fourth, according as parasitic connection might become ac-
centuated with herbaceous plants on one hand or with arbores-
cent plants on the other, in corresponding degree might at least
two lines of evolution open up; one in which entire plants para-
sitic on short-lived herbaceous hosts would become soft and de-
graded, while on the other hand, genera parasitic on roots of
woody and not least arborescent types, might become in corres-
ponding measure perennial and enlarged in the infesting region.

Detailed comparison will now be made.

General Morphology of the Root

In commencing parasitism as shown and illustrated in G. flava
(Plate XII, Fig. 1) the matured perennial primary root system
is spreading, loose, and expanded, covering a large area. From
this expanded primary root system, secondary rootlets pass out-
ward and downward and some of these end in parasitic suckers.
The roots are firm, strong, fibrous, and play no small part in col-
lecting raw material from the soil. The plant still depends large-
ly upon crude sap obtained directly from the soil for its susten-
ance, and could possibly live an independent existence as Kerner
(14, p. 180) has stated for Odontites.

In G. pedicularia the root system is more condensed and
suckers are more abundant.

In G. purpurea (Plate XII, Fig. 2) the secondary root system,
though delicate, is more condensed with numerous sucker en-
largements on a variety of plants such as Grasses, Composites,
etc. The main root from which these side rootlets arise is con-
densed and shortened.

From the short condensed primary root of G. aphylla (Plate
XII, Fig. 3) spring the secondary roots which are less spreading
than in the previous two species, wholly indicating a more close-
ly dependent parasitism on host roots. These are smaller than
the roots previously mentioned, being thin and fibrous with
abundant suckers.

Drawings of the above were made to illustrate natural size as
nearly as possible and to show the extent of primary and sec-
ondary roots.

In the transition from the above types the Scrophulariaceae
are connected with the Orobanchaceae by the genera Lathraea,

148 Boeshorc — The Morphological Continuity of

and Christisonia. The extent of the root system of Lathraea
is very limited; the roots are semi-woody semi-fibrous. This
genus, as already pointed out, is by some authors regarded as a
member of the Scrophulariaceae, by others as a member of the
Orobanchaceae. It offers a graded transition from pale green
and greenish purple to purple red genera like Buchnera, Har-
veya and Hyobanche of the Scrophulariaceae to Orobanche
of the Orobanchaceae. Kerner (15, p. 182) has described the
parasitism of Lathraea. He says: "The young root of the seed-
ling grows at first at the expense of reserve material stored in the
seed, penetrates vertically into the earth and sends out lateral
branches, which, like the main root, follow a serpentine course
and search in the loose damp earth for a suitable nutrient sub-
stratum. If one of these meets with a living root belonging to
an ash, poplar, hornbeam, hazel, or other angiospermous tree,
it fastens on to it at once and develops suckers at the points of
contact; these suckers are at first shaped like spherical buttons,
but soon acquire, as their size increases, the form of discs adher-
ent to the host's root by the flattened side and with the convex
hemispherical side turned toward the rootlet of the parasite.
These discoid suckers cling to the root attacked by means of a
viscid substance produced by the outermost layer of cells. As
in the case of the parasites already described, a bundle of ab-
sorption-cells grows out of the core of each sucker into the root
of the plant serving as the host, and the tips of the absorbent

cells reach to the wood of the root" "The

roots, which issued originally from the seedling, and their suckers
have long since ceased to meet the requirements in respect to
nourishment of so greatly augmented a structure, and therefore
additional adventitious roots are produced every year, spring-
ing from the stem and growing towards living woody branches
of the thickness of a finger, belonging to the root of the tree or
shrub that serves as host. When there, they bifurcate, forming
numerous thickish filiform arms, which lay themselves upon the
bark of the nutrient root and weave a regular web over it.
Sometimes two or three of these root filaments of the parasite
coalesce, forming tendrils, and the resemblance to a lace-work
or braid is then all the more pronounced. Suckers, such as
have been described, are developed by these root-filaments
laterally, and more especially on the ends of the branches."

Scrophulariaceae and Orobanchaceae 149

Did we know the parasitic root relations of Harveya and Ilyo-
banche,these in all probability form a more perfectly graded con-
nection between the two supposed distinct orders.

From the above we pass to Orobanche, e. g., 0. minor
(Plate XII, Fig. 4), in which a dense mass of secondary roots
starts from a swollen shortened primary-root and these roots
form by their surfaces intimate and close connections with the
host roots, as described and figured by Koch (16). In 0. cru-
enta the primary root swelling becomes more pronounced and
the place of junction between stem and root is collar-like in ap-
pearance. Here two lines of deviation seem to start in conden-
sing degradation.

One line, simplifying and short-lived, leads to Aphyllon
(Plate XIII, Fig. 6) in which as figured by the writer there is
even more close and extensive parasitic connection with the host
than in the previous types and the parasite itself, in that sec-
ondary and probably primary roots parasitize. At the base of
the short stem arises the primary root, now slightly swollen,
and from this are given off secondary roots in such numbers as
to form a very much tangled mass.

Another line leads to the larger stronger genus Epiphegus
(Plate XIII, Fig. 7) of annual duration, in which a tuberous
swelling from a half inch to an inch across represents a fused pri-
mary root below and a greatly condensed vegetative stem-axis
above. From the lower part or primary root short and now
functionless rootlets start, while from the condensed stem adven-
titious roots, similar in relation to the secondary ones, arise at
any point of the stem axis. Accordingly parasitic connection
with the host is easily and directly made by the germinating
primary root as already pointed out by Cooke and Schively (17).
The roots of Epiphegus form around and above the beech roots
on which this is parasitic. The secondary and adventitious
roots are short and delicate.

In Conopholis (Plate XVI, Fig. 33), which may be regarded
as the climax type of the group, the root system so far as known
is entirely hidden from external view, being represented by a
large swelling and which usually terminates the oak root on
which it grows. Alike roots and leaves have been entirely ab-
sorbed, although in Epiphegus rudiments of both are present.
From the swelling arise numerous flowering shoots.

150 Boeshore — The Morphological Continuity of

Further comparison of the last three genera will show that
the parasitic roots of Aphyllon do not cause truncation and de-
cay of the host roots, the latter remaining alive beyond the point
of attack, while in Epiphegus and Conopholis so complete is
the parasitism that host roots rarely remain alive beyond the
point of parasitic attachment. In almost all of the genera of
the Orobanchaceae parasitism has become highly specialized
with regard to selection of hosts.

Thus a continuous and easy gradation is traced from green
and nearly independent Scrophulariaceae to highly condensed
and degraded Orobanchaceae, which in Conopholis closely simu-
lates the most degraded representatives of the Balanophoraceae
and Rafflesiaceae. On the other hand, no even approximate or
suggested such connection is shown between Gesneraceae and
Orobanchaceae.

Histology of the Root

In all the species of Gerardia examined the internal structure
of the roots is of the radial polyarch type. The xylem is great-
ly in excess of the phloem and consists in young roots, like those
of G. purpurea, mainly of large pitted vessels; spiral tracheae
and xylem cells make up the rest of the xylem. The phloem
arms alternate with the xylem and are composed of the usual
elements. In older roots of G. purpurea, G. flava, G. aphylla,
the bundle system assumes a woody more dense character,
having the appearance of a ring of wood not unlike that seen in
a dicotyledonous stem, save for the presence of the phloem. A
considerable amount of hard bast develops. A several-layered
pericambium surrounds the bundle region and this in turn is
surrounded by an endodermal layer, easily recognized in young
roots. The cortex is variously developed: in small, young roots
it consists of cells in rows, of 2 to 3 cells in each, and these are
radially disposed and connect the epidermis with the other tis-
sues, while between these radial arms are very large open spaces;
in older roots the cortical space is filled with normal thin-walled
cells; in more mature roots of G. flava the cortex is composed
almost wholly of scleroid cells staining a deep red color in saf-
ranin. In mature roots the epidermis may be replaced by cork
tissue. Root hairs were also noted.

In the parasitism of Gerardia the roots produce swellings in
the regions of contact with the host, which become hemispher-

Scrophulariaceae and Orobanchaceae 151

ical and grow down the sides of the host roots but do not com-
pletely surround them. Sections of swellings show an epi-
dermis, a cortex, and bundle elements. The most conspicuous
feature of the bundles is the large number of cells with pitted
walls. The cells in the upper part of the swelling have not fused
to any great extent to form vessels, but in the lower spread-out
part of the swelling they have the appearance of vessels (with
pitted-reticulate thickenings in their walls) which establish a
connection with the xylem of the host roots. This was seen in
purpurea and pedicularia.

Roots of Harveya and Hyobanche were not available for inves-
tigation.

Transverse sections of the roots of Lathraea japonica show
the epidermis, a narrow cortex, and the polyarch bundle system.
In the latter unusually large pitted vessels make up the greater
part of the bundles.

The roots of Aphyllon are soft and delicate, drying quickly
when exposed to the air. They never completely surround the
host roots, nor do they develop as large swellings as seen in Ger-
ardia species when contact is made with the host. The orig-
inal root tissues seem to fuse more intimately with the corres-
ponding tissues of the host. Root hairs and root caps, as already
noted by Smith (18, p. 113), were not found.

The radial arch-system is maintained, but the bundles are few
in number; the small amount of xylem consisting of spiral tra-
cheae and pitted vessels is poorly developed. The amount of
phloem greatly exceeds that of the xylem, which is a feature we
should expect to find in a holoparasitic plant whose chief object
of parasitism is elaborated sap. The bundle system is surround-
ed by a pericambium and this in turn by a wide cortex of very
large cells with abundant quantities of starch. No stone cells
have been found in it. Surrounding the cortex is the epidermal
layer. So intimate and complete is the connection made with
the host roots that, epidermis of parasite is continuous with epi-
dermis of host, cortex of parasite with cortex of host, etc., and
were it not for the large cells of cortex with abundant contents
as compared with the same tissue of the host, the line of demar-
cation between the two could scarcely be distinguished. In all
this Aphyllon closely simulates species of Orobanche as figured
by Koch (19).

152 Boeshore — The Morphological Continuity of

As a still more condensed type Christisonia might well come
next to Aphyllon. This is described and figured by Worsdell
(20, p. 134). He describes three species of Christisonia, but
the one that concerns us here is Christisonia subacaulis. To
quote from him regarding it: "the most abnormal feature oc-
curring in these plants is presented by the subterranean portion
of Christisonia subacaulis especially, which, on investigation,
is discovered to consist of organs having the character of roots,
though their morphological nature is well concealed, owing to
their extreme modification arising from their parasitic habit
. . . The tubers, which arise at intervals in the root system
of the plant just named, are the most important parts of it, for
it is from these that the haustoria are chiefly formed, while they
also act as reservoirs of nutriment for the whole plant . . .
The haustorium is interesting as having an exogenous origin,
and not an endogenous one, as described for many other para-
sites; it agrees in this respect with that of Rhinanthus."

The roots of Epiphegus are shorter than the roots of Aphyllon
and seem to have lost all parasitic power, this function being
accomplished by the tuberous swelling. A transverse section
of such shows the host root deeply or shallowly buried within
the tissues of the tuber and thus connection is established be-
tween the two. In these degradation stages the tissues of the
parasite have become correspondingly more simplified as demon-
strated by Cooke and Schively (17).

In Conopholis (Plate XVI, Fig. 32) no external roots are visi-
ble. With the more complete and highly specialized paras-
itism has come a complete absorption of roots, or they may be
represented by a large mass of stone cells, but wholly buried
within the enormous swelling on the oak roots. One specimen
was found which measured 10 inches in length and 6 inches
across. Sections were made of smaller specimens but these
only showed masses of stone cells. All trace of definite systems
seem to have been lost. And so, Conopholis well represents,
both morphologically and physiologically, the climax of the en-
tire group.

In summing up the discussion on roots it can be said that, be-
ginning with G. flava and ending with the genus Lathraea, a
gradual condensation in extent of root system is accompanied
by a gradual degradation of root tissues which in Lathraea be-

Scrophulariaceae and Orobanchaceae 153

come semi-woody, semi-fibrous. With this degradation par-
asitism becomes of increasing importance. Further conden-
sation leads to Orobanche and Aphyllon with soft delicate roots
and such is accompanied by further simplification in root tis-
sues. From here the root system enlarges, becomes tuberous,
and is characterized by an increasing degree of hardness of the
tissues that reaches a climax in Conopholis. Further simplifi-
cation of tissues here results in unintelligible interpretation.

Comparative Morphology and Physiology of Stem and

Leaf

Under this heading stems and leaves will be considered. The
drawings accompanying this part of the paper were made to il-
lustrate more particularly relative lengths of stems.

In the most primitive Gerardias the stem is tall, well formed,
and typically normal dicotyledonous, attaining a height of 1 to
4 feet in G. flava (Plate XIV, Fig. 9). The lower leaves are
large, ovate-lanceolate, sinuate-toothed along the margins; in the
upper leaves the margins are entire. Both stem and leaves are
covered with a fine close down.

Transverse sections of the stem show a large pith area, a ring
of xylem with numerous pitted vessels and internal to these spir-
al tracheae, a cambium, a narrow zone of phloem consisting of
little soft and much hard bast. Externally the phloem is bound-
ed by a cortex zone of 1 to 5 layers of cells, with the outer region
collenchymatous and the inner region of thin-walled cells. The
epidermis is persistent around the stem and from it project the
numerous two to four-celled pointed hairs. The basal cell of
these is large, rounded; the outer cells are narrow with the tip
cell pointed.

G. purpurea, with stems 1 to 3 feet tall, has a comparatively
small pith, but a wide zone of xylem. The phloem consists
mainly of soft bast with occasional patches of hard bast. Scler-
enchyma cells are a frequent feature of the cortex. The leaves
of this species are much reduced in size, linear, acute, rough-
margined according to Gray (21, p. 731). The rough character
is due to many one-celled, pointed, more or less spiny hairs distri-
buted over the general leaf surface and especially along the leaf
margins.

154 Boeshore — The Morphological Continuity of

In G. aphylla (Plate XIV, Fig. 10) a very marked condensa-
tion occurs. The stem is slender, rather wiry, unbranched,
and from 6 to 18 inches in length. The leaves, much reduced in
size, are scale-like, tapering almost to a point, and closely ap-
plied to the stem. Stomata are distributed over the leaf surface
in comparatively small numbers when the entire foliage of the
plant is taken into account. Leaves and stem are covered with
short, conical, unicellular hairs.

Transverse sections of the stem taken at the same level as
in the preceding genera are smaller in circumference. Stems
tend to become quadrangular, being reinforced by scleroid cells
at the angles. A rather wide green cortex surrounds the bundle
system. Phloem consists of about equal amounts of hard and
soft bast. Xylem is well developed and wide, enclosing a small
pith area.

G. aspera (Plate XIV, Fig. n) shows further reduction in
height of stem. The internal structure is very similar to that of
G. aphylla. Stem and leaves have hairs similar to those of G.
purpurea.

Condensation advances even more markedly in Harveya and
Hyobanche (Plate XIV, Fig. 12, 13) in that the stem is only from
4 to 6 inches in height. Both genera are parasitic, have scale-like
leaves or these even reduced to functionless scales. Wettstein
(22,p.97) describes Hyobanche as a fleshy low, parasitic plant with
numerous scale-formed leaves, the lower scales being smaller
than the upper ones. Material of this genus was not available
for study, but the foregoing description regarding the fleshy na-
ture of the plant suggests a very strong tendency toward genera
of Orobanchaceae.

Transverse sections of the stem of H. coccinea have a wide
pith around which the ring of bundles is arranged. These are
more simplified than in the preceding genera, and in this respect
Harveya is a good transition from species of Gerardia to Oroban-
che of the Orobanchaceae, in fact it is closer to Orobanche than to
Gerardia. The xylem, while forming a continuous ring, is nar-
row in places and irregular; its elements are spiral tracheae and
pitted vessels. The phloem is in excess of the xylem, soft bast
of 1 to 3 layers of cells in patches, and hard bast that is very strik-
ingly developed as 3 to 6 layers of cells forming a solid ring about
the soft bast. The cortex is as wide as the phloem and xylem

Scrophulariaceae and Orobanchaceae 155

combined and in places attains twice their extent in width; the
cell walls show irregularity. The epidermis, from herbarium
material, could not be described very accurately.

The hairs of H. coccinea are of special interest because of
their large number, great size, and capitate-glandular character.
They are distributed over both stem and scales; as many as 20
were counted around the edge of a cross section of the stem of
reasonable thickness, which does not account for any that were
broken off. The "stalk" of each hair consists of from 3 to 5
cells; the basal cell is usually short, broad, and rounded, while
the others are elongated almost cylindrical; where two cells join,
there is a collar-like constriction, giving the "stalk" the appear-
ance of being jointed. The glandular tip is composed of 2 to 4
cells. Here again this genus is very similar to Orobanche which
will be described later.

Proceeding from the last mentioned genus through the genera
of Orobanchaceae the vegetative axes become more or less in-
conspicuous though enlarged fleshy (in some genera) and tuber-
ous, and partly or wholly subterranean.

In the transition genus Lathraea (Plate XIII, Fig. 14) the
vegetative axis has become a rhizome from 2 to 4 or 6 inches
long, and is provided with the hollow scale leaves concerning
which so much discussion has taken place, while from the an-
nual short-lived inflorescence numerous flowers arise. The
plant is described by Kerner (23, p. 135) as being destitute of
chlorophyll. "The subterranean stems are white, have a fleshy,
solid, and elastic appearance, and are covered throughout their
entire length with thick squamous leaves placed closely one
above the other. " The leaves are broadly cordate. The scales
being underground, naturally have lost their vegetative func-
tion, but are provided with cavities and structures for catching
animal prey. Solereder (24, p. 586) says that the function as-
cribed by Kerner and Wettstein to these structures is incorrect
according to Scherffel and Heinricher. Kerner describes two
kinds of structures formed on the internal surfaces of the scale
for which he suggests no special name in Lathraea squamaria;
others have called them glands. One kind is composed of a cy-
lindrical stalk cell and two cells forming a head, which project
into the cavity of the scale; the other variety, which does not
project into the scale, is composed of a tabular cell and two con-

156 Boeshore — The Morphological Continuity of

vex cells, forming a low dome amongst the epidermal cells. The
writer examined scales of L japonica and found similar struc-
tures projecting into the cavities of the scale. The cells of the
epidermal tissues are from 4- to 6-sided and regular in outline.
Stomata are either absent or present in small numbers; struc-
tures were found that had much the appearance of stomata, but
no accurate statement regarding them could be made.

In the genus Orobanche (Plate XIII, Fig. 15) the vegetative
axis is an inch and a half or less in length. If the rhizome of
Lathraea be thought of as shortened and as a consequence the
axis became tuberous and enlarged, the condition of affairs in
Orobanche would be reached in which the stem axis is a short-
ened but enlarged tuber, covered densely with crowded scales.
These are more elongated than the scales of Lathraea, e. g.,
in 0. minor, broad oval at the base and lanceolate in their upper
part; along the flowering axis they are lanceolate.

In 0. cruenta the stem is exceptionally broad, measuring three-
fourths of an inch across.

In transverse sections of the stem of 0. minor their outline
is found to be very irregular, which is due to some extent to the
scales given off at different levels. There is however, irregular-
ity also due to grooves and rather wide ridges, which become
greater in number and more pronounced on the floral axis.
Such sections show sections of the scales, as well, from whose
outer surface are given off the glandular hairs, which agree in
description with the hairs of Harveya. From 3 to 5 cells form
a stalk with the capitate part of the hair of 2 to 4 cells. The
cortex is unmodified, consisting of large thin-walled cells which
are packed with starch grains. The bundle system consists of
bundles arranged in an irregular manner about the pith area;
phloem is next to the cortex and xylem next to the pith. The
pith area is large and like the cortex has considerable starch in
its cells.

Transverse sections of the flowering axis of 0. coerulea show
the irregularity in outline mentioned above to the greatest de-
gree. The epidermal cells have heavy outer walls. The hairs
are of the same type as for Harveya and 0. minor, and are
quite numerous. The cortex cells are thin-walled and have less
starch than those of the stem cortex. The bundle system tends
to become more loose and open than in the previous genera and

Scrophulariaceae and Orobanchaceae 157

in places the bundles are separated; the amount of phloem is
about twice that of the xylem and consists of hard and soft bast;
the xylem has a few spiral tracheae and pitted vessels, that we
best see in longitudinal sections.

The vegetative axis of Aphyllon (Plate XIII, Fig. 16) is
reduced in thickness to scarcely more than one-fourth of an inch,
or usually less, but in length is about one inch. The number of
scale leaves is reduced to from 5 to 10, alternately placed and
separated on the stem; they are smaller than the scales of Oro-
banche and decrease in size from the upper to the lower ones.
The lower ones have neither stomata nor hairs, while the upper
ones have both. The hairs are capitate and multicellular, like
those of Harvey a and Orobanche.

In making cross sections of the stem their soft and fleshy nat-
ure is at once recognized. The epidermis consists of small cells
which are thickened on the free side; amongst these, stomata
may be seen. The most conspicuous feature of the stem is the
unusually large rounded thin-walled cells in the cortex and pith.
Both regions are wide in extent and their cells are packed with
large starch grains. Between these regions a comparatively
narrow ring of bundles is arranged. The bundles are more wide-
ly separated than in Orobanche, thus making the medullary
rays quite wide. Around the bundle system is a sheath of from
2 to 4 layers of cells, much smaller than the cortex cells and
whose greater diameter is placed in a tangential direction. The
excess amount of phloem over xylem is more pronounced than
in the previous two genera. The phloem is external and xylem
internal, the latter composed of a few spiral tracheae and pitted-
reticulate cells.

Transverse sections of the flower stalk show the same arrange-
ment of tissues as in the stem. The bundles are placed in a nar-
row ring about the pith and internal to the cortex; the xylem is
slightly better developed than in the stem, while the phloem
forms a continuous zone about the xylem. Cortex and pith
have some starch grains. The epidermis, of small cells with
free walls heavily thickened, is frequently interrupted by cells
in groups of two slightly raised above its surface which appear
to be stomata. Glandular capitate hairs, arising in great num-
ber from the flower stalk, have also two cells forming their base.

In Epiphegus (Plate XIII, Fig. 7) the vegetative stem and
enlarged primary root tubercle become confluent into an oval

158 Boeshore — The Morphological Continuity 0}

or rounded tuber from three-fourths to an inch and a half long.
The writer considers the upper part of the tuber as the stem axis
which bears the tooth-like scale-leaves. In some specimens the
stem part consists of more than half of the tuber, in others of
less than half. The scales are shorter, but more numerous than
those of Aphyllon. Stomata are present on the scales, and
along the scale-edge a few multicellular hairs were seen. Ad-
ventitious roots arise from the surface of the stem.

Histological details have been worked out by Cooke and
Schively and the writer has found his own investigations to agree
with their descriptions. But by way of comparison of this genus
with preceding genera some of these details will bear repetition.
The most notable feature is the bundle system, which becomes
broken up into separate bundles that have no definite arrange-
ment in rings, thus giving the stem a more loose and open as-
pect. This is an advance in degradation from Aphyllon in
which the bundles, though separated, form a fairly definite ring
about the pith. The phloem greatly exceeds the xylem, a fur-
ther advance in degradation with increased parasitic habit. To
quote shortly from Cooke and Schively: "The phloem of a
bundle .... shows a tendency to spread out and lie in
separate patches, while the xylem of each bundle seems always
concentrated in a single area. Many of the bundles show an in-
ternal duplication with reversed order, phloem, xylem, xylem,
phloem, succeeding each other from without inwards .
An internal phloem is almost always present, often in excess of
the outer phloem mass." No such state of affairs as the last
exists in Aphyllon. As a consequence of the scattered and
elongated bundles the cortex and pith areas have become re-
duced in size.

Sections of the floral axis have been described by others. The
bundles are arranged in a ring. Considerable hard bast is de-
veloped. Hairs and stomata constitute the epidermal growths,
the former showing considerable reduction in size as compared
with the hairs on the floral axis of Aphyllon.

No definite statement will be ventured as to how much of the
plant constitutes the stem axis in Conopholis (Plate XVI, Fig.
32). The aerial portion in its lower part is densely covered with
scales that may represent reduced and degraded leaves, while
slightly above this point the scales become more distinct, sep-

Scrophulariaceae and Orobanchaceae 159

arated from one another, and increased in size, averaging 4 to
6 times the size of the lower ones. Sections made in the region
of the crowded small scales show numerous patches of hard
scleroid cells and the inner almost continuous ring of bundles,
while in the position of the outer ring some fairly recognizable
bundles are seen. Above this point the two rings of bundles
become complete and the scleroid patches disappear; below this
point the bundles become fewer in number and less recogniz-
able, but the scleroid patches increase in number until in the
tuberous part scleroid patches are the predominant feature and
the bundles as such disappear. Evidently the lower part of the
aerial portion represents a transition from stem axis to aerial
shoot, and so, this part together with the upper part of the tub-
erous swelling might be considered as the stem axis. As in
Epiphegus, stem system and root system have become con-
fluent with no sharp line of distinction between the two.

Stomata have been reported absent from the scales but sev-
eral were found on the lower or outer surface of the upper scales.
From their shape they appear to be almost or quite function-
less (Plate XV, Fig. 28). They are misshapen and poorly de-
veloped; some have two elongated guard cells which have
slipped out of position and show a long orifice between them,
others have three and four guard cells very loosely fitted to-
gether. Large multicellular and unicellular hairs are found on
the edges of the scales which the drawings (Plate XV, Fig. 31)
will sufficiently indicate.

The histology of the flowering axis is interesting for several
reasons. The cortex and pith areas are quite large and consist
of rounded cells of varying size and thickness with frequent
large intercellular spaces. Some of the cells have contents while
others are empty. Two definite rings of bundles surround the
pith, which have a zone of fundamental tissue intervening.
This quite agrees with Wilson's observations (25, p. 14) but dis-
agrees with those of Chatin (26, p. 590) who says that Epiphegus
and Conopholis have three rings. The majority of bundles
in both rings are completely separated from each other by fun-
damental tissue, and this represents the climax in the series,
beginning with Harveya in which the bundle system shows no
tendency toward separation of bundles, thence to Orobanche
which shows indications of such, thence to Aphyllon in which

160 Boeshore — The Morphological Continuity of

the bundles show slight separation, thence to Epiphegus in
which the bundle system is loose and broken up into separate
bundles, which finally in Conopholis become still more widely
separated and distinct.

As to the elements making up the bundles, the following is
stated by Wilson (25): "Each bundle of the inner row has in-
ternally xylem, made of xylem cells and well-developed spiral
tracheae. Next to the xylem is found the phloem, which in a
longitudinal section proves to consist of both sieve tubes and
companion cells. Adjacent to the phloem are a number of par-
enchyma cells, whose walls are so angular and so much thick-
ened that in the photograph these bundles appear to be bi-col-
lateral. That such is not the case, however, is easily proved on
longitudinal section, when the parenchymatous nature of these
cells is at once visible. Even in cross section, the color of the
walls differentiates the wood from the thickened parenchyma.

The bundles of the exterior row have the same structure as
those of the interior, only the xylem is now exterior so that the
phloem masses of the two rows face each other."

Sections stained in safranin and methyl green also some in
Delafield's haematoxylin and safranin bring out the bundle ele-
ments quite plainly. A different interpretation from the above,
however, is suggested with reference to the arrangement of the
elements of the bundle. Each bundle of the inner ring shows
internally phloem, consisting of a large patch of hard bast and
a patch of soft bast. The xylem comes next and is poorly de-
veloped, consisting of an interrupted line of cells, which are al-
most wholly large spiral tracheae, running across the bundle;
in some bundles there are several rows of spiral tracheae and
amongst them are several cells of phloem. Beyond the xylem
is another small patch of soft bast and next to it another small
patch of hard bast. The bundles of the outer ring show the
same elements, but in reverse order, small patches of hard and
soft bast, xylem, large patches of soft and hard bast.

Comparative Study of the Inflorescence

Regarding the inflorescence of the two supposed distinct fam-
ilies, several points are of special interest. In the parasitic
Scrophulariaceae the flowering axes are elevated above the sur-
face of the soil by the more or less elongated vegetative axes in

Scrophulariaceae and Orobanchaceae 161

the less parasitic forms, but as these intervening vegetative axes
become shortened and increasingly degraded in the more par-
asitic forms like Harveya and Hyobanche the flowering stalks
gradually approach a lower level until they take their origin
only a few inches from the ground. The conspicuous part of
the plant then consists of the short vegetative axis and the floral
axis. In the Orobanchaceae with the greatly reduced stems the
flowering axes constitute almost wholly the aerial parts of the
plant. In the former the flowering axes are woody, while in
the latter they are more or less fleshy and stout, with one excep-
tion (Aphyllon).

The type of inflorescence for the Scrophulariaceae is given
by systematists as centripetal, racemose. Gerardia flava has
one floral axis along which short-pedicelled flowers are arranged
racemosely. In G. purpurea, G. aspera, G. aphylla and
others there are several floral axes, each of which constitutes a
raceme; the pedicels are of varying lengths. In some of the
more parasitic forms the branching of the flower stalk is less fre-
quent and the flowers have very short pedicels so that the type
of inflorescence tends to become a loose spike, e. g. Orthocarpus
purpurea, Euphrasia, Bartsia viscosa, Odontites, Hyobanche, etc.

Aphyllon has one floral axis bearing a terminal flower. Epip-
hegus has branched flowering stalks and the flowers are ar-
ranged racemosely or in a spike. In the other genera, Lath-
raea, Orobanche, and Conopholis, of the Orobanchaceae the
flowers are arranged in a dense spike.

The bracts of the flowers in Scrophulariaceae are leaf -like
becoming scale-like, or even scales in Hyobanche while in Oro-
banchaceae they are scales.

Comparative Study of the Flower

Calyx.

The calyx in the different genera studied shows considerable
variation in the number of sepals, in form, and in hairs.

In species of Pedicularis the calyx may be in the form of a
funnel, or bell-shaped with from 2 to 5 teeth at the top. The
teeth may be simple or further toothed or lobed.

The calyx of Melampyrum lineare is comparatively small and
is made up of a short tube below and 4 tapering teeth above, the

1 62 Boeshore — The Morphological Continuity of

teeth being longer than the tube. These are in pairs, with the
upper pair slightly longer than the lower pair. Short, blunt,
unicellular hairs and stomata are distributed over the surface
of the calyx.

In Bellardia the calyx is much larger than in Melampyrum,
and consists of the lower half of a bell-shape while the upper
half is divided into 4 teeth. As in Melampyrum, the two lower
teeth are about two-thirds the size of the upper ones. A prom-
inent vein runs into each tooth. The hairs are numerous and of
two kinds; very long, slender, unicellular, pointed hairs are
found mainly along the edges of the teeth and prominent veins,
and the less numerous 3- to 4-celled glandular hairs only along
the main veins and edges of the teeth ; while between the prom-
inent veins they are almost wholly of the pointed type, but short
and unicellular. Stomata are also present.

In Fistularia (Alectorolophus Bieb., Rhinanthus L. p. p.) the
calyx consists of 4 sepals in pairs, the parts of which are united
almost to their tips, but the pairs are separated for more than
half of their length. Two sides of the calyx are pressed together
in the young state, which in the fruiting stage become inflated
and persistent around the fruit. Two types of hair are found
on the surface; a short unicellular, and a long multicellular
pointed type, also a glandular type with two rounded cells form-
ing the top.

In species of Gerardia the bell-shaped calyx is 5-toothed, the
teeth in most cases are shorter than the tube. In G. purpurea the
teeth are very short to half the length of the tube and sharp-
pointed. G. flava was referred to before as covered with a fine
close down, which character applies to the calyx as well. A
peculiar feature of the hairs in this species is the spiral thicken-
ings found in them. The short pointed type is rare or wanting
entirely. In G. aspera and G. purpurea the short, pointed hairs
are the predominant type; these are both one- and several-celled.
The latter has also short glandular hairs. Stomata are a com-
mon feature in the three species. The epidermal cells are reg-
ular in outline.

In Euphrasia americana the calyx is small, 4-toothed, the
teeth being longer than the tube and lanceolate. Pointed uni-
cellular hairs are quite common, but are not found all the way to
the tip of the calyx teeth. The short glandular hairs are con-

Scrophulariaceae and Orobanchaceae 163

fined largely to the angles formed by the bases of the calyx teeth.
The former type is roughened along the margins. Epider-
mal cell walls are wavy.

Bartsia alpina has a relatively large calyx of 4 parts, with the
teeth about as long as the tubular part. Long-stalked gland-
ular hairs are abundant; the stalk consists of 3 to 6 slender cells,
while the top of it has from 2 to 5 cells arranged in parallel rows.
Pointed multicellular hairs, though present, are few in number.
Epidermal cells are irregular in outline.

In Harveya the tubular part of the calyx is short and the 5
lobes extend almost to the base. As already described in con-
nection with the stem histology, the hairs are of maximum size
and wholly glandular. The stalk cells are comparatively wide
and stout, are 3 to 6 in number, and capped by the gland cells.
Unlike the majority of the genera already described, the hairs
in this genus fringe the edges of the sepals.

Hyobanche has a calyy of 5 parts which are almost distinct
and more rounded than in Harveya.

Among the Orobanchaceae similar variation may be seen in
the calyx, as in Scrophulariaceae. According to Le Maout and
Decaisne (27, p. 593) the "calyx is persistent, tubular or cam-
panulate, 4-5 fid, or of 4 sepals more or less completely united in
lateral pairs."

In species of Orobanche variation occurs, for the calyx may be
split both above and below, nearly or quite to the base; the di-
visions may be 2-cleft or entire, or more or less unequally 2- to
5-toothed. In 0. coerulea and 0. minor the teeth are about as
long as the tube, and are lanceolate-subulate. The hairs on
these two species are similar to those of Harveya but smaller.

The calyx of Epiphegus is small, the teeth very much shorter
than the tube below. Cooke and Schively (28, p. 377) state
the following concerning hairs: "One-celled, rarely two-celled
hairs fringe the edges of the lobes. Below, across the base of
the lobes, there extends a band of two or three-celled hairs,
longer than the UDper hairs. All of these hairs are on the outer
surface of the calyx; none are present on the inner surface. They
have a swollen granular appearance." The writer has found
his own descriptions to agree with the above.

In Lathraea the calyx is bell-shaped with 4 to 5 rounded teeth
above.

164 Boeshore — The Morphological Continuity of

The calyx of Aphyllon is 5-toothed, the teeth equal to, or
longer than the tube. Glandular hairs are numerous and dis-
tributed over the entire outer surface of the calyx; shortly, these
consist of a short broad basal cell, then a long cylindric cell,
next to this 2 to 3 cells decreasing in size, and finally the top or
capitate part of several rounded cells. Stomata are numerous
and of the normal type.

In Conopholis the calyx is orbicular, split in front, and toothed
at the tip. Hairs and stomata are negligible in this genus.

Stamens.

The stamens in parasitic Scrophulariaceae and Orobanchaceae
are 4, didynamous, in Gesneraceae, 5 to 4 to 2.

As to anthers, a totally different relation holds between par-
asitic Scrophulariaceae and Orobanchaceae on the one hand as
compared with Gesneraceae on the other. In the former the
top of the filament is broadly inserted into the swollen back of
the anther, the lobes of which in the parasitic Scrophulariaceae
and Orobanchaceae are prolonged above the connections, but
in a striking manner are prolonged downward as two parallel
or divergent awns. These anther lobes in all except two gen-
era of Scrophulariaceae (Harveya, Hyobanche) and two genera
of Orobanchaceae (Campbellia, Aeginetia) are like each other,
that is, are equally paired anther lobes. But in Harveya one
anther lobe is large, normal, and polleniferous, and prolonged
below into a long horn; the other lobe is small, abortive, and
radiates back from the top of the filament. The latter anther
lobe in Hyobanche has been entirely absorbed, so that now one
fertile anther lobe dehisces by a single basilar pore.

It is of interest to find that in Christisonia and Aeginetia of
the Orobanchaceae a similar structure exists, for, as in Hyo-
banche and Harveya, each bears a single fertile anther lobe.

As to antherine structure, this exactly agrees in parasitic
Scrophulariaceae and Orobanchaceae and totally differs from
anything encountered in Gesneraceae. The structure and ap-
pearance of such genera as Melampyrum, Tozzia, Euphrasia,
Pedicularis, Buchnera, Gerardia, Bartsia, etc., absolutely resemble
those of Orobanchaceae, and this the writer would regard as
one of the most important points of contact between the two
families.

Scrophulariaceae and Orobanchaceae 165

The anthers in the various genera of the Gesneraceae con-
form to a totally different type from that shown alike in Scro-
phulariaceae and Orobanchaceae. In the former, when the
flowers become nearly or quite regular as in cultivated varieties
of Gloxinia and also in species of Streptocarpus, as well as in other
genera, the anther lobes are adpressed at their tips and cohere
so as to form a solid antherine box. When the flowers are ir-
regular, as in nearly all members of the family, the 4 anthers are
pressed together in pairs by their tips. In the entire group fur-
ther, the bases of the anthers are either parallel and with blunt
rounded extremities, or the anther lobes diverge in their lower
part but are blunt and rounded in their divaricate bases. In
no case studied by the writer is there any indication of the anther
lobes being arranged in parallel pairs that at their lower ex-
tremities become prolonged into horned or horn-like awns.
This peculiarity is typical, as will be traced, in most of the
known parasitic Scrophulariaceae and directly continued as a
character to the Orobanchaceae.

Of all the genera examined Gerardia flava might well be placed
at the top of the list regarding the size of its anthers. As illus-
trated in Plate XV, Fig. 17, the anther lobes are large, ellip-
tical, rounded at the upper extremities and at their lower ex-
tremities are projected abruptly into two long tapered, some-
what divergent processes. The front and sides of the anthers
are covered with numerous long, multicellular hairs. These
are present also on the filaments.

The anthers of G. purpurea (Plate XV, Fig. 18) are smaller and
taper more gradually downward into processes that lie almost
parallel with each other. The filament is inserted above the
middle of the anther lobes. Anthers and filaments are hairy.

In G. aphylla (Plate XV, Fig. 24) and G. aspera the processes'
are relatively short and less pointed than in the former species.
Both anthers and filaments are covered with hairs.

In Melampyrum lineare the anthers agree fairly well with those
of G. purpurea, but are smaller. Along the filaments knob-like
swellings of 1 to 2 cells are seen.

The antherine condition of Harveya is set forth in Plate XV,
Fig. 19 and has already been referred to.

Other genera of Scrophulariaceae like Bartsia (Plate XV, Fig.
20) Tozzia, Bellardia, show likewise the downwardly-direc-
ted anther-processes.

1 66 Boeshore — The Morphological Continuity of

For Lathraea squamaria Kerner figures basal anther-processes
similar to those in the above genera.

The stamens of two species of Orobanche, also those of Aphyl-
lon, Epiphegus and Conopholis are figured in Plate XV, Figs.
25, 26, 27, 23, 22, all of which have the downwardly-directed
processes.

In Orobanche coerulea the anther lobes are round above and
taper gradually downward into rather short processes at their
lower extremities, which are slightly convergent.

In 0. minor the shape of the anther lobes is approximately
rectangular, with the processes given off from the sides next to
the filament.

In Aphyllon the anthers are comparatively small. The fila-
ment is inserted above the middle of the anther lobes, with the
processes at their lower extremities.

Hairs on the stamens are a constant feature in all the above.

The parasitic Scrophulariaceae and Orobanchaceae thus have
a character in common. Further, this character becomes more
important when the function of the processes is considered, for
in both groups these are contrivances designed to aid in the shed-
ding and dissemination of the pollen grains. With regard to
the genus Bartsia, Knuth (29, p. 229) says of B. apula the fol-
lowing: "In this Dalmatian species each anther possesses a
downwardly-pointing process, which is pushed to one side by
insects, thus opening the pollen receptacle and causing pollen
to be sprinkled on the head and back of the visitor."

In connection with the description of pollination in Lathraea
squamaria the statement is made by Knuth that the pollen can
not fall out until the short, blunt point of an anther receives a
blow from an insect. Again, under species of Orobanche the
following is stated: "The four anthers are laterally united, and
each lobe is provided with a sharp, stiff, downwardly-directed
process. These processes are behind the stigma, and if any-
thing strikes against them the bright-yellow, powdery pollen
falls out of the anther-lobes, and is sprinkled on the proboscis
or head of the visitor."

Histologically the structure of the awns shows a striking
agreement throughout the genera of Scrophulariaceae and Oro-
banchaceae already mentioned, in that as one passes from each
anther lobe toward their downwardly-directed awns the epider-

Scrophulariaceae and Orobanchaceae 167

mis or exothecial tissues become increasingly thickened on their
outer and lateral walls in u-shaped manner until toward the tip
of each awn the thickening may be almost as deep as the cell
cavity. Transverse and longitudinal sections of the awns of
Aphyllon are figured in Plate XV, Figs. 29, 30.

Were the above not a continuous morphological series from
green to degraded parasitic plants, one can scarcely suppose
that such similarities and histological details could have evolved
in two such related families.

Pistil

In a study of the pistil the fundamental point for consider-
ation is the supposed invariable two-celled ovary in Scrophulari-
aceae and the one-celled ovary in Orobanchaceae and Gesner-
aceae. This has already been discussed generally on pages 6-8.
The varying structural details from a two- to one-celled con-
dition, as well as the diverse views expressed as to the affinities
of such genera as Hyobanche, Lathraea, Christisonia that show
wavering transition from two- to one-celled states, emphasize
again the fact that here we are dealing with a condensing and
simplifying variation. No such transition-relations are even
suggested between Gesneraceae and Orobanchaceae.

In the different genera of the Orobanchaceae the fused mar-
gins of the carpels grow inward to a varying degree from shallow
marginal placentas to deep parietal, that approach central pla-
centation. So the fundamental point of supposed affinities be-
tween Gesneraceae and Orobanchaceae entirely breaks down,
while a natural and continuous affinity between degraded par-
asitic Scrophulariaceae and still more degraded Orobanchaceae
has been established.

Hairs of the Pistil

The presence of hairs on the style and ovary in the different
genera of the two families is not a constant feature. They differ
as to number, distribution, and type. The following results
were noted:

In Gerardia aspera the hairs on the style, though present, are
few in number.

In species of Odontites the base of the style and upper part
of the ovary are very hairy, while the upper part of the style is
almost glabrous. The hairs are long, narrow and pointed.

1 68 Boeshore — The Morphological Continuity of

In Rhinanthus the hairs are numerous from the stigma down-
ward but decrease in number toward the base of the style and
finally disappear. They consist of one cell pointed at the dis-
tal end on the upper part of the style, and of 2 to 3 cells on the
lower part of the style, and are broad at the base, tapering grad-
ually toward the tip.

In Bartsia alpina the numerous one- to several-celled hairs
are distributed along the entire style and the upper half of the
ovarian surface, being most numerous on the ovary. They are
quite long and needle-like in appearance.

In Bellardia they are exceedingly numerous along the entire
stylar surface, and consist of one needle-like cell that is dark in
color.

In Orobanche coerulea the glandular type of hair is seen along
the entire style, similar to the hairs found on other parts of the
plant already described.

The style and ovary in the genera Aphyllon, Epiphegus and
Conopholis are glabrous, although the stigmatic areas of the first
two mentioned are covered with comparatively short unicellu-
lar hairs.

The Nectary

The nectarv of Scrophulariaceae is described by Wettstein
(30, p. 39) as hypogynous, ring formed or one-sided. Beck (31,
p. 127) gives a similar description for Orobanchaceae. "Nektar
absondernden Stellen am Grunde der Stf. oder am Grunde des
Frkn. ringformige, oft buckelig, seltener beutelformig vorspring-
ende Nektarien. " For Gesneraceae Fritsch describes the nectary
as a "Discus" usuallv well developed, ring to cup-shaped, or as
reduced isolated glands, which may be also one-sided. While
these general descriptions fairly agree in the three families, most-
ly one-sided nectaries are characteristic of parasitic Scrophulari-
aceae and Orobanchaceae, as will now be taken up.

In Melampyrum pratense "the nectary expands toward the
lower lip into a whitish rounded body, on either side of which
runs a nectar-secreting groove."

In M. lineare the nectary consists of a similarly rounded body
placed to one side of the ovary.

The general description of the nectary for Rhinanthus is given
by Knuth. He says that nectar is secreted by the fleshy base

Scrophulariaceae and Orobanchaceae 169

of the ovary which projects to the front, and it is stored in the
bottom of the corolla tube. In all the species examined by the
writer the nectary agrees with this description, but further, it
is curved inward at the top and points directly toward the ovary
as a tongue-like process.

In Bartsia alpina the nectar is secreted by a cushion-like
swelling at the lower side of the base of the ovary, extending a
little beyond as a rounded knob.

Lathraea squamaria has a large, roundly triangular and some-
what lobed nectary situated at the base of the ovary.

In L. clandestina the ovary is laterally compressed and tra-
versed bv a longitudinal groove, bearing in front a three-lobed
nectary.

One species of Orobanche, crenata, "secretes nectar at the
orange-yellow base of the ovary."

The nectary of Aphyllon is a small, rounded, whitish swelling
at the base and a little to one side of the ovary.

In Epiphegus the nectary appears as a swelling on one side
of the ovarv, antero-laterally in position, just above the base.

The nectary of Conopholis is a rudimentary ovarian gland.

The Seeds

The seeds vary in number from 4 in one capsule of some gen-
era of the Scrophulariaceae, (Melampyrum, Rhinanthus) to very
numerous (as many as 1500 in one capsule) in other genera of
Scrophulariaceae and Orobanchaceae. Increase in seed number
is usually accompanied by a reduction in size, so that the seeds
become very small in such genera as Epiphegus and Aphyllon.
In structural details, the seeds of the purely parasitic genera are
simplified and most degraded; in some of them the embryo con-
sists of a small group of undifferentiated cells {Aphyllon, Epi-
phegus.)

To supplement the writer's information, gathered from his
own examination of material, Bentham and Hooker's "Genera
Plantarum" (32, pp. 967-980) was used as the chief source for
genera of Scrophulariaceae.

The seeds of Rhinanthus are few in number, sub-orbicular,
compressed, and surrounded by a wing-like structure. The
embryo is small.

Melampyrum has from 2 to 4 seeds which are smooth, and have
an aril-like appendage at the base.

170 Boeshore — The Morphological Continuity of

In Gerardia, the seeds are numerous, oblong wedge-shaped
or angular; the testa is loose-fitting.

In Euphrasia, the numerous seeds are pendulous, oblong,
with longitudinal ridges.

In Bartsia, the seeds are many, pendulous, or may be numer-
ous and placed subtransversely, with wings and longitudinal
ridges.

In Tozzia, the seeds are ovoid-globose; the testa is appressed;
and the embryo is small.

The seeds of Buchnera are very numerous and ovoid or oblong;
the testa is reticulate, subappressed.

In Harveya, the seeds are very numerous; the testa is heavily
reticulated and loose; the embryo is equal to half the albumen in
amount.

In Hyobanche, the seeds are numerous, small, globose; the
testa is loose, reticulated.

Lathraea has numerous small seeds, spherical in shape, and
the testa is wrinkled.

In species of Orobanche, the seeds are numerous, reticulated,
wrinkled or striate; the embryo is minute with cotyledons scarce-
ly differentiated.

In Christisonia, the seeds are extremely numerous, very small,
and subglobose, and the testa is reticulated.

In Aphyllon, the seeds are numerous, small, light, surrounded
by a tough leathery coat of flattened cells with thick indurated
walls; the endosperm cells are filled with starch and enclose a
small embryo consisting of a group of undifferentiated cells.

In Conopholis, the seeds are of fair size, numerous, and some-
what quadrangular in shape; the embryo is small, undifferent-
iated; and the testa is heavily thickened.
t In Epiphegus, the seeds are very numerous (from 700 to 1800)
small, oblong in shape; the embryo is a group of undifferentiated
cells; and the testa cells are elongated with much thickened
walls.

Selection of Hosts

The less parasitic Scrophulariaceae have a rather wide range
of hosts. The species of Gerardia parasitize on Grasses, Com-
posites, etc., as has already been stated.

For Bartsia, the following are given as hosts: Avena flaves-
cens, Phleum pratense, Trifolium pratense.

Scrophulariaceae and Orobanchaceae 171

Tozzia parasitizes on the roots of Ranunculus, Petasites, Rumex,
and Alchemilla.

Poa, Avena, Luzula, Carex, Senecio, Trifolium, Capsella, Epi-
lobium, and Festuca, are given by Heinricher as hosts of Euph-
rasia.

The roots of Lathraea may attach themselves to the roots of
Ash, Elm, Poplar, Hornbeam, and Hazel as hosts.

Species of Orobanche seem to have a very wide range of hosts.
According to Koch (16), Orobanche minor may parasitize on 44
species of plants; 0. ramosa on 29 species; 0. speciosa on 13
species, and 0. hederae on 3 species. These hosts may be found
amongst members of Papilionaceae, Geraniaceae, Cruciferae,
Oleaceae, Ranunculaceae, and others.

Although various species have been given as hosts of Aphyllon,
the writer quite agrees with Smith in finding it to grow only on
the roots of Aster corymbosum.

Epiphegus has been found to grow only on the roots of Fagus
americana. Conopholis similarly parasitizes on but the one
genus Quercus, and so far as the writer has learned on the group
of the red oaks.

The writer feels deeply indebted to Professor John M. Mac-
farlane who first suggested the work and whose valuable criti-
cisms, and assistance have been a source of constant encour-
agement in the preparation of this paper. Thanks are due to
Mr. W. R. Taylor for the photographs in Plate XVI, to Dr. F.
W. Pennell for the use of several slides, and to Mr. H. W. Stout
for locating a growth of Conopholis.

Summary

A short review of the evidences dealt with above might be
put in summary form as follows:

1. All macroscopic and microscopic details suggest that the
parasitic Scrophulariaceae and Orobanchaceae form a contin-
uous and parasitically degrading morphological series that show
transitional steps from green nearly autotrophic plants like
Melampyrum, Rhinanthus, and Euphrasia to increasingly con-
densed and degraded genera like Bartsia and Harveya, on to
Lathraea, that has been shown to be placed by some botanists
in Scrophulariaceae, by others in Orobanchaceae, thence through
species of Orobanche to Epiphegus, and finally Aphyllon and

172 Boeshore — The Morphological Continuity of

Conopholis. No such continuity exists between the wholly
green autotrophic Gesneraceae and Orobanchaceae, nor are any
members of the Gesneraceae parasitic.

2. During progressive parasitism in Scrophulariaceae and
Orobanchaceae commencing parasitism consists in a few of the
fibrous roots becoming enlarged toward their extremities into
parasitic haustoria, while other roots are still autotrophic in
relation. With increasing parasitism these secondary roots be-
come shortened and the primary root also condenses into a cen-
tral knob or swelling as can be traced successively in Gerardia,
Lathraea, Orobanche, Epiphegus, Conopholis and Aphyllon.
Ultimately, complete vegetative fusion and enlargements be-
tween primary root and ascending vegetative axis result in the
formation of a rounded tuber {Epiphegus) or greatly swollen
rounded or oval mass {Conopholis) difficultly distinguishable
from the enlarged roots of oak on which the last grows.

3. In progressive degradation the elongated ascending leafy
axis of Gerardia or Bartsia shortens steadily, and in Harvey a and
Hyobanche becomes a short axis bearing reduced nonchloro-
phylloid foliage leaves. These are seen in Lathraea to become
the characteristic scales investigated by numerous observers.
In Orobanche the leafy axis is reduced to a short tuber that is
separated from the condensed primary root by a constricted
neck {0. cruenta), or the primary root and scaly axis become
continuous as in 0. minor. This by progressive degradation
becomes a slightly constricted root part below and a stem part
above, covered by tooth-shaped leaf-scales, or an oval tuber on-
ly, that is primary root below and tooth-scale stem above {Epi-
phegus). In Conopholis even this distinction is largely oblit-
erated.

4. In the less parasitic types, e. g., G. flava, the leaves are
large green and actively vegetative, but by gradual stages be-
come in time small and scale-like in G. aphylla. In Harveya
and Hyobanche they are scattered along 2 to 6 inches of the con-
densing axis. In Lathraea, these scales are largely underground
colorless, or purplish-white, and extend over 1 to 3 inches of the
vegetative shoot. In Orobanche the brown, yellow or red scales
cover the short tuberous vegetative axis for one to one-fourth
of an inch. The same is true for Epiphegus, or the scales are
rather shorter for Aphyllon. In Conopholis the intimate fusion

Scrophulariaceae and Orobanchaceae 173

of root and vegetative axis has resulted in practical obliteration
of any distinction of parts.

5. The inflorescence axis from being greatly elongated to
constitute a several- to many-flowered raceme becomes a rela-
tively simplified axis with a reduced number of flowers, the cli-
max of which is reached in Conopholis, which bears a spike with
40 to 10 flowers. In Aphyllon each inflorescence is represented
by a single flower.

6. It has been shown that the histological details, of stem and
leaf in the above progressive series of degrading parasitic types,
agree fundamentally with and verify the naked eye characters.

7. The sepals of Scrophulariaceae and Orobanchaceae have
been found to show fundamentally similar structures and at
times to show similar condensation in the former from a five-
leaved to a four-leaved calyx through absorption of the odd se-
pal, as in Tozzia, Euphrasia and others. Lathraea, as an inter-
mediate type, may have a five to four-lobed calyx, and finally,
species of Orobanche and Boschniakia may show five to three se-
pals making up the calyx.

8. In structure the stamens have been shown to constitute
an important link in the chain of evidence, for in all Gesneraceae
the anther lobes more or less converge and press against each
other at the apex and are rounded, as well as often divaricate at
their bases. In Scrophulariaceae and Orobanchaceae the mac-
roscopic and microscopic details proclaim progressive modifi-
cations in that the bases of the parallel anther lobes grow down-
ward into stiff awn-like horns, whose terminal cells are similarly
thickened throughout the series, and show a similar mode of
unequal thickening. In pollination, therefore, authors like
Ogle, Miiller, and Knuth demonstrate similar pollination ar-
rangements for Scrophulariaceae and Orobanchaceae that differ
markedly from those in Gesneraceae.

9. Histologically, the sepals and stamens have been shown in
many of the types to bear tapered multicellular hairs inter-
spersed with capitate-glandular hairs that suggest strongly an
origin in common.

10. The nectary in Gesneraceae is a cylindric structure that
appears either as a simple girdle, or as a series of connected
nectariferous knobs; very rarely is it a median unpaired swell-
ing. In Scrophulariaceae and Orobanchaceae the nectary is

174 Boeshore — The Morphological Continuity of

often a median knob in line with the antero-posterior axis of the
flower, or somewhat displaced.

1 1 . Evidence has been given to show that while the ovary is
two-celled in Scrophulariaceae and usually one-celled with deep
to shallow placentas in Orobanchaceae, all transitions between
these can be traced from Harvey a and Hyobanche to Lathraea
clandestine and Christisonia albida, etc.

12. As to seeds, these are few (4-12) and fairly large in the
less parasitic Scrophulariaceae, becoming decidedly small and
numerous in Harveya and allied types of Scrophulariaceae as
well as in all of the Orobanchaceae.

13. In structural details and morphological complexity the
seeds show continuous degradation changes from the less para-
sitic genera of Scrophulariaceae to the most degraded genera of
Orobanchaceae, such as Aphyilon and Conopholis, in which a
rudimentary endosperm and formless embryo are alone devel-
oped up to the time of germination.

14. Physiologically it has been shown that in transition from
the less parasitic Scrophulariaceae, like Odontites and Gerardia
which parasitize on a variety of hosts, transitions are shown to
genera like Orobanche, some species of which seem to confine
their parasitism to the species of a family or even to the species
of a genus, as pointed out by Beck (p. 31), while finally in such
a highly degraded type as Epiphegus abundant evidence shows
it to be purely parasitic on Fagus americana, Aphyilon to be
similarly wholly parasitic on Aster corymbosum, and Conopholis
on one or two species of Quercus.

Conclusions

From a review of the above observations, the writer believes
that ample evidence has been adduced to show that direct and
distinct continuity can be established from non-parasitic through
semi-parasitic Scrophulariaceae to the most degraded parasites
of the family, and that these again show direct continuity with
the still more degraded and condensedly parasitic types of Oro-
banchaceae.

Alike logically and biologically, therefore, the two types
should be treated in continuous descending series from the high-
est to the most degraded genera.

Scrophulariaceae and Orobanchaceae 175

Explanation of Plates.

Key to the lettering of Figures 1-16.

r — The primary root from which are given off secondary rootlets.

st — Stem.

hr — Host root

fl — Foliage leaf.

fs — Foliage scale.

i — Inflorescence axis.

Plate XII.

Fig. 1. Root system of Gerardia flava showing parasitic attachment.
(From a drawing by Mr. J. Stauffer in Gray's "Structural Botany.")
Fig. 2. Root system of Gerardia purpurea. Natural size.
Fig. 3. Root system of Gerardia aphylla. Natural size.
Fig. 4. Root system of Orobanche minor.

Plate XIII.

Fig. 5. Root system and part of stem of Orobanche minor attached to

clover root. (After Koch).
Fig. 6. Stem and root system of Aphyllon uniflorum parasitic on the

roots of Aster corymbosum. Natural size.
Fig. 7. Tuberous swelling consisting of stem and root system of Epi-

phegus virginiana. Natural size.
Fig. 8. Stem and root system of Orobanche cruenta with constricted neck

at the junction of stem and root. Natural size.
Fig. 14. Root, underground stem, and lower part of inflorescence axis

of Lathraea japonica. (X }4).
Fig. 15. Orobanche minor. Note the short condensed stem, from i

downward, as compared with that of Lathraea. (X }4).
Fig. 16. Aphyllon. Stem still more condensed and shorter than that

of Orobanche. (X }4).

Plate XIV — Note the gradual condensation and shortening of stem in Figs.
9-13. Foliage leaves become scales.
Fig. 9. Gerardia flava. (X Vi)-
Fig. 10. Gerardia aphylla. (X /€)•
Fig. 11. Gerardia aspera. (X XA).
Fig. 12. Harveya capensis. (X ]4)-
Fig. 13. Hyobanche. (X 14).

Plate XV — Figures 17-27 show the downwardly-directed processes at the
base of the anther lobes. (X 10).

Fig. 17. Stamen of Gerardia flava.

Fig. 18. Stamen of Gerardia purpurea.

Fig. 19. Stamen of Harveya coccinea. Note the elongated sterile an-
ther lobe.

Fig. 20. Stamen of Bartsia.

Fig. 21. Stamen of Melampyrum.

176 Boeshore — The Morphological Continuity of

Fig. 22. Stamen of Conopholis, front and back views.

Fig. 23. Stamen of Epiphegus, front and back views.

Fig. 24. Stamen of Gerardia aphylla.

Fig. 25. Stamen of Orobanche minor.

Fig. 26. Stamen of Orobanche coerulea.

Fig. 27. Stamen of Aphyllon nniflorum, front and back views.

Fig. 28. Epidermal cells and stomata on a scale of Conopholis. (X

ii5).

Fig. 29. Transverse section of the awn-like horn at the base of an an-
ther of Aphyllon. The epidermal tissue shows cells with the outer and
radial walls much thickened in u-shaped manner. (X 1 15)-

Fig. 30. A longitudinal section of the awn-like horn of an anther of
Aphyllon. (X 115).

Plate XVI.

Fig. 32. Aerial flowering shoots of Conopholis showing attachment at
the base to the swelling on the oak root. Originally there were 40
shoots growing from the one swelling. (X }4).

Fig- 33- Tuberous swelling on oak root showing the numerous excres-
cences from which the flowering shoots arise. (X H)-

LITERATURE CITED

1. Wettstein, R. Die naturlichen Pflanzenfamilien. IV. 3. B, 1895.

2. Beck, G. R. Die naturlichen Pflanzenfamilien, IV. 3. B, 1895.

3. Fritsch, Karl. Die naturlichen Pflanzenfamilien IV. 3. B, 1895.

4. Baillon, H. Histoire des Plantes, X, 1891.

5. Le Maout, E. and Decaisne, J. Descriptive and Analytical Botany,
English Translation, 1876.

6. Warming, E. A Handbook of Systematic Botany, 1895.

7. Heinricher, E. Die griinen Halbschmarotzer. Jahrbiicher fur wis-
senschaftliche Botanik, 32, 1898.

Die griinen Halbschmarotzer. Jahrbiicher fur wissenschaftliche Botanik,

36, 1901.

8. Solms-Laubach, H. Ueber den Bau u. die Entwicklung parasitischer
Phanerogamen. Jahrbiicher fur wissenschaftliche Botanik, 6, 1867-68.

9. Beck, G. R. Monographie der Gattung Orobanche. Bibliotheca Bot-
anica, Heft 19, 1890.

10. Koch, L. Die Entwicklungsgeschichte der Orobanchen, 1887.

11. Henderson, M. W. Contributions from the Botanical Laboratory
of the University of Pennsylvania, V, 1919.

12. Beck, G. R. Die naturlichen Pflanzenfamilien, IV. 3. B, 1895.

13. Worsdell, W. C. On the Comparative Anatomy of Certain Spec-
ies of the Genus Christisonia. Annals of Botany. IX, 1895.

14. Kerner, A. The Natural History of Plants, translated by Oliver,

L 1895-

15. Kerner, A. The Natural History of Plants, translated by Oliver,

I, 1895.

16. Koch, L. Die Entwicklungsgeschichte der Orobanchen, 1887.

Scropludariaceae and Orobanchaceae 177

1 7. Cooke, E. and Schively, A. F. Observations on the Structure and De-
velopment of Epiphegus virginiana. Contributions from the Botanical Lab-
oratory of the University of Pennsylvania, II, No. I, 1904.

18. Smith, A. C. The Structure and Parasitism of Aphyllonuniflorum,
Gray. Contributions from the Botanical Laboratory of the University of
Pennsylvania, II, No. I, 1904.

19. Koch, L. Die Entwicklungsgeschichte der Orobanchen, 1887.

20. Worsdell, W. C. On the Comparative Anatomy of Certain Spec-
ies of the Genus Christisonia. Annals of Botany, IX, 1895.

21. Gray, A. New Manual of Botany, Seventh Ed., 1908.

22. Wettstein, R. Die natiirlichen Pflanzenfamilien, IV, 3. B, 1895.

23. Kerner, A. The Natural History of Plants, translated by Oliver,

L 1895-

24. Solereder, H. Systematic Anatomy of the Dicotyledons, trans-
lated by Boodle and Fritsch, I, 1908.

25. Wilson, L. L. W. Observations on Conopholis Americana, Contri-
butions from the Botanical Laboratory of the University of Pennsylvania, II
No. 1, 1904.

26. Chatin, G. A. Translation in Solereder's "Systematic Anatomy of
the Dicotyledons," I, 1908.

27. Le Maout, E. and Decaisne, J. Descriptive and Analytical Bot-
any, English translation, 1876.

28. Cooke, E. and Schively, A. F. Observations on the Structure and
Development of Ephiphegus virginiana. Contributions from the Botanical
Laboratory of the University of Pennsylvania, II. No. I, 1904.

29. Knuth, P. Handbook of Flower Pollination, III, 1909.

30. Wettstein, R. Die natiirlichen Pflanzenfamilien, IV, 3. B, 1895.

31. Beck, G. Die natiirlichen Pflanzenfamilien, IV, 3. B, 1895.

32. Bentham, G. and Hooker, J. D. Genera Plantarum, II, 1876.

Bol. Contrib. Univ. Perm.

Vol. V, Plate VI.

Taylor on Reproduction in Acer.

Bot. Contrib. Univ. Penn

Vol. V, Plate VII.

Taylor on Reproduction in Acer.

Bot. Contrib. Univ. Penn. Vol. V, Plate IX.

.••• •• • • • •

•••• •: •

z ,

•a
51

•

„-. -.:• &% •>• •.•;•

52 53 54 55

56

57

/V*" 59 61 62 60 -\* I •

63 * 64

#

• «.

#•*•.*

t #

•'V

72 73 •*

76 77 78

Taylor on Reproduction in Acer.

Bat. Contrib. Univ. Penn.

Vol. V, Plate X.

79

«• • •

• • • • • *

• • ••• •

•••••••

81 .• •••.••

•••■••

•••• 82

• •

/..v.

v • ••

• ••

• • • •»• •

• •••v

Taylor on Reproduction in Acer.

Bot. Con t rib. Univ. Pom.

Vol. V, Plate XI.

105 ^-^__^106 108

Taylor on Reproduction in Acer.

Bot. Contrib. Univ. Pom.

Vol. V, Plate XII.

BOESHORE ON SCPOPHULARIACEAE AND OROJ3ANCHACEAE

Bot. Con I rib. Univ. Penn.

Vol. V, Plate XIII.

BOESHORE ON SCROPHULARIACEAE AND OROBANCHACEAE

Bol. Contrib. Univ. Penn.

Vol. V, Plate XIV.

<f

&

^

BOESHORE ON SCROPHULARIACEAE AND OROBANCHACEAE

hot. Contrib. Univ. Penn.

Vol. V, Plate XV.

25

26 M / / 27

BOESHORE ON SCROPHULARIACEAE AND OROBANCHACEAE

Bot. Contrib. Univ. Pemi.

Vol. V. Plate XVI.

32

33

BOESHORE ON SCROPHULARIACEAE AND OROBANCHACEAE

3 5185 00293 :