The Structure and Life-History

of the

Hay-Scented Fern.

BY

HENRY SHOEMAKER CONARD, Pn. D.

WASHINGTON, D. C.
PUBLISHED BY THE CARNEGIE INSTITUTION OF WASHINGTON.

1908.

CARNEGIE INSTITUTION OF WASHINGTON
PUBLICATION No. 94

THK COKNMAN IMMNTINi; CD.,
CARLISLE, 1'A.

PREFACE.

Aside from certain personal interests and opinions, the impulse to the
present investigation came from a study of recent papers by Jeffrey, Boodle,
and Gwynne-Vaug'han. But since we shall never know the true relations
of a plant to its surrounding's until we have worked out its complete life-
history, it seemed to me very desirable to have all of our knowledge of
this species collected into a unit. Therefore the study was carried beyond
the problems suggested by the papers referred to.

The work was begun in odd moments of an instructorship at the Uni-
versity of Pennsylvania, but nearly all of it was actually done in the
Botanical Laboratory of the Johns Hopkins University, and this paper is to
be regarded as Contribution No. 7 from that Laboratory. 1 was there as
"James Buchanan Johnston Scholar" from February, 1905, until June,
1906. For the opportunity to carry on this investigation in a peculiarly
stimulating atmosphere, I am deeply indebted to those who administer
the affairs of the university. It is an especial pleasure to express appre-
ciation of the constant friendly interest taken by Prof. Duncan S. Johnson.

„ V.

The fundamental teachings of Prof. W. K. Brooks have also molded many
of my thoughts and expressions. Thanks are due to Mr. I. F. Lewis fora
collection of material from Long Island; to the late Mr. E. R. Heacock for
my first pot of prothallia and "sporelings;" to Dr. C. E. Waters for infor-
mation and for the excellent photographs, plates 1 and 2; to Henry Holt &
Co. for the use of two copyrighted pictures; to Capt. John Donnell Smith
for library facilities; to Mr. J. D. Thompson, of the Library of Congress,
and Mr. Joseph H. Painter and Mr. W. R. Maxon, of the United States
National Museum, for looking up certain papers not otherwise accessible
to me; and to the officers of the Academy of Natural Sciences of Philadel-
phia for the use of several rare old books. All of these obligations are now
gratefully acknowledged .

HENRY S. CONARD.
GRINNELI,, IOWA, April, 1907.

THE STRUCTURE AND LIFE-HISTORY OF THE HAY-SCENTED FERN.

By HENRY SHOEMAKER CONAKD,
Professor of Botany, Iowa College.

HISTORICAL INTRODUCTION.

The hay-scented fern, Dcnnsticdtia fyutictilobida (Michx.) Moore (=
Dicksonia futnctilobnla Willd.) first appeared in botanical literature in 1803,
when it was described by Michanx as follows:

\Nephrodiuni\ punctilobuliini. N. majusculum: stipite nudo, ramis pinnulisque pu-
bescentibus: fronde longa, bipinnata; pinnulis decurrentibus, subovali-oblongis, semi et
ultra pinnatifidis; lobulis oblonguisculi, apice 2-4-dentatis, singulis unipunctiferis. Obs.:
Habitus Polyp, filicis fojinina- Linn. Hab. in Canada. [A. Michaux, 1803, p. 268.]

There is nothing in the text to indicate that this is a new species.
Michaux' s genus Ncphrodiuiii was extremely far-reaching", being- defined
in these words: "fructibus punctis subreniformibus" (p. 266). Among
the species are Ar. thclyptcroides, marginale, filix-ftzmina, and dryoptcris!

Swartz (1806) placed the hay-scented fern in the genus Aspidium, in
which he was followed by Willdenow (1810). The latter writer, both in
his own text and in his quotation from Michaux, changes the spelling of
the specific name to funictilobiim. But he had already (1809) described it
under the name of Dicksotiia pilosiuscula, and this, too, is copied in the
Species Plantarum. The text of the Enumeratio (1809) is as follows:

DICKSONIA.

Sori subrotundi distinct! marginales. Indusium duplex, alterum superficiarum exte-
rius dehiscens, alterum e margine frondis inflexo interius dehiscens.

i. DICKSONIA pilosiuscula.

D. frondibus bipinnatis, pinnis pinnatifidis, laciniis dentatis, rachi pilosiuscula.
Polypodium pilosiusculum. Miihlenberg in litt. Habitat in Pennsylvania. (!)
'4 D.

An important addition to the other diagnosis is the notice of hairs upon
the rachis. These are so characteristic as readily to distinguish this fern
f n >m any other in our native flora. In preparing the ' ' Species' ' , Willdenow
recognized the similarity of his Aspidium punctilobum and Dicksonia pi lost its -
cula as expressed in the closing words of the description of the latter :

An Aspidium punctilobum supra p. 270 dubie indicatum, eadem sit filix aliis ad
dijudicandum relinquo? quum pinnule neque sint decurrentes neque pubescentes.

6 STRUCTURE AND LIFE-HISTORY OF HAY-SCENTED FERN.

Schkuhr (1809, p. 125, plate 131) referred to this fern as Dicksonia
t)i(t>fsccns* He has been followed only by Presl (1836, p. 136).

Desvaux (1827) made this species the type and only member of his
genus Sitoboliiun. His diagnosis of the genus reads: "Sori globosi; in-
volucrtmi fornicatum globulosum a basi ad apicem dehiseens" (pp. 262,
263). No specific diagnosis is given. t J. Smith (1841) changed the
spelling to Sitolobhan, and Newman's text (1851) gives Litolobium. G.
Kunze writes thus in Linneea (23:249): Sitoboliuin (mafeSitoIobium),"
but in 1848 the printer makes him say " ' Litolobiiim {not Sitolobimu )."
Link's (1841 ) genus . ldectum\ is too late ever to be more than a synonym.
The identity of the plant, however, has never been in doubt, for it stands
absolutely unique amid its native surroundings. The list of synonymy
on page 45 will serve to show how the name has been bowled about.

Its generic affinities are briefly discussed on page 42. \Vc will simply
state that its place is at present established in Bernhardi's (1800) genus
Dfnnst(cdlici (type: D. Jlaccida - Trichomanes flaccidum Forst.), and we

*On plate 131 marked Dicksonia pubescens. Text on p. 125 reads:

II. DICKS, [pubescens in margin of page] frondibus subtripinnatis, foliolis lanceo-
latis, pinnis oblongis, laciniis ovatis dentatis, stipite glabro, rachi pubescente. Siv.
Mohr. in Lift.

Nephr odium punctilobuhim, maiusculum; stipite nudo, ramis pinnulisque pubescen-
tibus: fronde longa, bipinnata; pinnulis decurrentibus, subovali-oblongis, semi et ultra
pinnatifidis; lobulis oblongiusculis, apice 2-4-dentatis, singulis unipunctiferis. Mich.
Flor. Bor. Ainer. n. p. 268.

Habitat in Canada. Habitus Polypod. filic. fern. Mich.

Weichhaariger Dicksonischer Fani. Mit fast 3-mal getiedertem Laube, lanzet-
formigen Blattern, langlichen Blattchen, eyrunden, gezahnten Lappen, glatten Strunke
und eine weichhaarigen Spindel.

Dieser Farn erhielt ich stiickweise aus Pennsylvanien auch voter Polypodiutn pilo-
siuscitluin Willd., wonach ich zwar die eigendliche Grosse nicht, aber nach dessen
Theilen doch die 3-fache Fiederung erkennen kann. . . . [The next paragraph de-
scribes the plate, closing with the words] Alle Rippen der Blattchen and Lappen sind,
wie die Spindel, mil gegliederten Haaren bekleidet.

tDesvaux's full text reads:

SITOBOLIUM N. Sori globosi; involucrum fornicatum globulosum a basi ad apicem
dehiseens.

i. ^S". punctilobum N. Nephrodium punctilobum Mich., Fl. am. bar., 11. p. 268.
Aspid. punctilobum S\v., -Sj;/., p. 60. Dicksonia pilosiuscula Willd., En. Jiort. her., p.
1076. Dickson. pubescens Schk., Fil., t. 131.

J Link's full text is as follows:

ADECTUM.

Frons tripinnatisecta. Sori subrotundi marginales ad sinus frondis. Indnsium
undique ad sorum adnatum eumque tegens, dernum medio dehiseens et circulare.

A. Dicksonia defectu sporidochii valde differt.

i. A. pilosiusciilum fr. tripinnatifida, pinnellis brevibus, antice et superne incisis,
stipite rhachi costisque pubescent! bus.

D. Fr. 1-2 ped. alta, pinnae 3 poll. Igae., pinnulae 3 lin. Igae.

Dicksonia pilosiuscula W. sp. 484. IV. E. 1076. E. a. 2.464. H. b. 2.10. Raddi
bras. 6^,. Dicksonia pubescens Schkuhr kr . 125 /. 132.

Hab. in sylvis opacis ad rupes Pennsylvaniae et Virginiae nee non in locis montosis
prope Tejuco Brisiliae. Perenne. [p. 72.]

HISTORICAL INTRODUCTION. 7

follow Moore (1857) and most recent scholars in accepting the name Denn-
sttedtia f>u 11 cfilolni la (Miehx.) Moore.

Two varieties of /). frioniilobula have been described in recent years.
Dennsteedtia punctilobula V2X . cristata Maxon (1899) was found in Massa-
chusetts by F. G. Floyd. Under cultivation the percentage of crested
fronds produced varies greatly. "Some fronds have not only had the
apex of every pinna doubly or trebly crested, but the apex of the frond
itself has frequently been bifidly divided with heavily crested apices"
(Davenport, 1905). I have several times seen fronds with the rachis bifur-
cated 10 cm. or more below the apex. Each fork, then, bears a normal
continuation of the leaf. Waters (1903) considers this condition "fre-
quent." He also states (p. 289) that "A form with rather narrow fronds,
the pinnae unequal in length and with the teeth of the ultimate segments
very deeply cut, so that each vein forms the midrib of a narrow tongue-
like segment, has been named D. fii/o$iitsci/la schizophylla" Of course this
name should read Dennstadtia punctilobula schizophylla. On the relation
of these varieties to the typical form I have no opinion to express.

In botanical literature other than taxonomic or floristic the hay-scented
fern scarcely appears. Descriptions of its habit of growth, its glands, and
long, slender rhizome are given by Williamson (1878, p. 117, plates XLV,
XLVI), Eaton (1879-1880, pp. 341-343, plate 44), Clute (1901, pp. 225-231),
and Waters (1903, pp. 288-290). Frances Wilson writes an appreciative
general account of these features in the Asa Cray Bulletin (1897), and
Waters (1903) adds to a pleasing text photographs which are exquisite
and true to life. Parsons (1899) and Eastman (1904) refer to this fern in
a popular way.

Eaton (1879-1880) and Waters (1903) speak of the concentric arrange-
ment of light and dark tissues in the rhizome (</". fig. 67), and the tax-
onomic writers tell of the indusium in detail. De Bary (1884) describes
the vascular bundle of Dcnnstiediia (naming this species along with three
others in parenthesis) as having a tubular bundle, "which is closed a;s far
as the foliar gap; the bundle which enters the leaf arises from the whole
margin of the gap as a continuous concave plate \_cf. fig. 82] , which is
only occasionally split up at its base into several bundles lying side by
side." Gwymic-Yaughan in 1901 (p. 85) mentioned the present species as
showing thick-walled elements in the phloem of the petiole. In 1903
(p. 691) he includes D. punctilolnila in a list of nineteen ferns with typ-
ical and practically identical solenostelic structure, as described by him
in Loxosoma ciiuninghaiini. A page of text is devoted to a summary of
the facts of structure in the group. A summary of taxonomic literature,
with synonym)-, is given on our pages 44 and 45 following.

8

STRUCTURE AND LIFE-HISTORY OF HAY-SCENTED FERN.

SPOROPHYTE.

The hay-scented fern occurs generally in open woods (fig', l) or clear-
ings or on roadside banks. It prefers well-drained, stony or sandy soil,
and usually forms thick beds. In the Catskill Mountains of New York
and in New England it grows on the cleared hillsides in dense patches 8
to 15 m. in diameter. Its range is from New Brunswick to Alabama and
Minnesota (Britton and Brown, 1896, 1 :12). The leaves are from 50 to 90
cm. high, lanceolate, and thrice pinnatifid. A light-green color and dense
pubescence combine to give the fern a soft, feathery appearance. The
glandular hairs exhale a delicate fragrance when brushed, which has been
likened to new-mown hay; hence the common name.* The stems are found
5 to 15 cm. beneath the surface of the soil — long, slender, much-branching"
rhizomes (fig. 3). These spread rapidly from year to year, and give rise
to the densely matted beds of the fern. Roots of threadlike fineness arise
plentifully from all parts of the rhizomes and ramify through the soil.

THE ROOT.

The roots are numerous, cylindrical, with copious, two-ranked branching.
They extend more horizontally than vertically in the soil, and do not descend
below 20 cm. from the surface. The color is black in mature portions,
shading" in the younger parts through reddish-brown and brownish-yellow
to creamy white at the apex. Although but 0.5 mm. in diameter (maximum
0.545 mm.; minimum 0.49 mm.; average 0.523 mm.), they are tough and
wiry in texture. The rootlets (secondary roots) are about half as thick as
the main roots. Tertiary roots, similar to the secondary, frequently occur.
Only rarely does a root arise from the base of a leaf, and then it is usually
within 4 mm. of the center of the rhizome.

TABLE i. — Acropetal development of roots from stem.

Length
of root.

Distance
from apex
of stem.

Branching.

Collected.

nnn.
A 77

nun.
•3 . 2

N o branch es

Univ. of Pa., 8/6/"o4.

6 4O

J
•J . 2

No branches

Fallsington, Pa., io/4/'o3.

3-2

4.78

J -

4-5
6.40

One branch 4.7 mm. long, i
cm. from stem.
Many branches

Do.
Do.

From any part of the stem roots may come out, but more frequently
from the lower side. A stem 5 cm. long, including the tip, showed eleven
roots, inserted as shown in fig. 9. They arise in acropetal succession very

*The names fine-haired fern, mountain fern, gossamer fern, and hairy Dicksonia are
given by Clute (1901, p. 231), and sweet grass fern by Eastman (1904, p. 67).

vSPOROPHYTE. 9

near to the stem apex (see table l) and lengthen rapidly. I have one
which is 24.5 cm. long- and the broken ends were frequently found at a
distance of 20 cm. from the stem. It is likely that these lengths are not
much exceeded. The rootlets are generally alternate on opposite sides of
the primary root-axis, but many exceptions occur. Two rootlets are often

TABLE 2.

found consecutively on one side, and in
one case five were seen. Successive root-
lets may be as much as 8 mm. apart or
almost or quite opposite (fig's. 6, 7, 8;
table 2). None occur usually within 12
or 15 mm. of the stem.

Table 2 shows irregularities in alter-
nate arrangement of rootlets on opposite
sides of roots. The figures indicate
distance in millimeters of each rootlet
from the one next above it , and columns
show alternation.

The root-cap is rather long and pointed
(fig-. 23). From the initial cell of the
root outward ten rows of cells may be
seen in a strongly developed specimen,
five in a slender root of a sporeling.
Outside of these cell-layers a worn-out
layer is seen, in the act of sloughing- off .
The inner layers are small-celled and
rich in protoplasm. The outermost cells
are four times the diameter of the inner,
but still nucleated. Indeed, small, dense,
nucleolus-like nuclei are seen even in the
layer that is shedding". The cap thins
out layer by layer along the sides of the
root, and the cells become very long and
narrow. The outermost layer persists
some distance above the next inner one.
No sign of statolith bodies has been
seen in any part of the root-cap.

In development, each terminal segment of the root-initial gives rise to
a single layer of root-cap cells.* The segment divides first by an anti-
clinal wall parallel to one of the sides of the initial (figs. 10, 11, 26). In
successive cap segments the first wrall of one stands either directly over or
at an angle of 60° to that of the preceding or following one, and not at

Root 1.

Root 3.

Right.

Left.

Right,

Left.

Stem

Stem

13-5

13.0

3-5

3-5

1.6

2.O

3-2

2 .O

3-5

I .5

1.8

3-3

3-5

3-5

3-0

3 -°

4.0

i . 5

6.2

8.0

4.0

3-3

O.2

4-5

4-o

6-5

2 .O

3-5

6.5

Koot 2.

3-°

4-0

(?1

2 .O

/

4.0

5-5

6.0

2-5

5.0

0.5

3-5

I . I

3-o

4.0

3-3

t-5

Root 4.

4-4

6-5

3-5

(?)

5 -°

2-5

1.6

2-5

6-5

1.8

5-0

3.0

8.0

*In a few cases periclinal walls were seen in three to five or six of the median cells,
making the segment two-layered at that point (fig. 24).

10 STRUCTURE AND LIFE-HISTORY OF HAY-SCENTED FERN.

45° , as stated by Nageli and Leitgeb (1868, p. 76). The halves are next
cut into quadrants by anticlinals at rig-lit angles to the first wall (fig's. 11,
12, 25). The succeeding- walls in the quadrants are heterodromous and
may be parallel to either of the preceding or be oblique (figs. 13, 25).
No further regularity was found in the division of root-cap segments.

The initial cell of the root is a triangular pyramid with its longest axis
in the axis of the root (figs. 23, 27, 28). Lateral segments are cut off
around the initial on one side after another in regular order. I noted ten
roots (fig. 27) in which the succession of segments was counter-clockwise
(proceeding from older to younger segments) and four in which it was
clockwise, as one views the cell from its outer (capward) base. Each
lateral segment divides first by a periclinal wall near its outer margin
(figs. 14, 15, 23, 27-29). The next wall is a radial anticline which passes
inward from near the middle of the first, and strikes one of the sides of
the segment near its inner angle, dividing the inner cell into "sextants"
(fig. 16, n; segment 3 in figs. 27-29). Thus there is in each segment a
larger (major) and a smaller (minor) sextant. In transverse section of
the root we see the three major sextants meeting at the center of the sec-
tion (figs. 29-33), with three alternating minor sextants which do not
reach quite to the center. The "sextant wall" meets that side of the seg-
ment which is adjacent to the next older segments, (kathodic wall) and is
therefore katadromous. As all of the segments in any root are alike in
this respect, the divisions are said to be homodromous. Soon after the
sextant wall is formed in the inner part of the segment it is laid down
also in the outer part (fig. 17; segment 4 in figs. 27-29).

A second pericline is now laid down near the middle of each inner
sextant cell (figs. 18, 23; segment 5 in figs. 27-30). As this wall forms
the boundary between the plerome and outer tissues, it may be called the
ficrifilcromc wall. Another periclinal laid down in the two outer sexants
divides these into two layers, the definitive epidermis (piliferous layer)
and hypodermis (fig. 19; segment 6 in figs. 27-31). Both of these tissues
remain one-layered throughout. Subsequent divisions in them are all
anticlinal, either radial or transverse (figs. 23; 27-33). Almost simul-
taneously periclinal walls are formed on each side of the periplerome wall,
near and parallel to it (figs. 14, 20, 21, -walls vi and vn; 23). In the
majority of cases, however, the outer one seems to precede. The result-
ing cells constitute the definitive endodermis and pericycle. The cells of
the former are from the beginning flattened, of the latter nearly cubical
(figs. 23, 31, 33).

If we group the segments into cycles of three, beginning with the latest
formed (</. figs. 27, 28), we find walls I and n (figs. 14-22) already in one
or more of the youngest cycle. Walls n, in, and iv are found in the
second cycle, and vi and vn in the second or third. In the second or

SPOROPHYTE. 11

*•

third cycle of segments longitudinal radial anticlines are also formed in
the outer members of the segment, dividing the sextants into halves (figs.
21, 22, vin). As yet each segment consists of hnt one layer of cells.
Transverse anticlines occurring throughout the segment in the third or
fourth cycle make it two-layered (fig. 23). A second series of divisions
in the same plane cuts the outer tissues (epidermis, hypodermis, cortex)
into four vertical layers. These cleavages occur first in the epidermis,
but their order of sequence is rapid and apparently varied. The pericyclc
is late in becoming divided; and the endodermis of a certain pair of oppo-
site sextants lacks the radial division for a long time, as will be described
in speaking of the origin of lateral rootlets.

The large cell remaining between the endodermis and hypodermis (figs.
14, 20, 21, 23) gives rise to all of the cortex. After its transverse anti-
clinal division it rapidly undergoes one to three periclinal and as many
radial divisions. The periclinals probably take place in centrifugal order.
The result is a cortex of two to four concentric layers, each with 14 to 24
cells. The last divisions are complete in the fourth, or at least the fifth
cycle of segments ((/". figs. 28-33).

The triangular cells lying within the pericyclc (fig. 21) divide either
by a periclinal wall into two parts or by two tangential walls into three
parts (figs. 22, 29-31). The tips of the three major sextants and of one
minor (occasionally two) become tracheids of the metaxylem. The.cells
between the tracheids and the pericycle of this minor and of the major
opposite to it become protoxylem cells (figs. 14, 22, 29-33); the inter-
mediate parts of the other two majors, and all within the pericycle in the
two remaining minors (with the exception noted above) go to form phloem
(figs. 14, 22, 29-33).

As the elements elongate, the transverse limits of the segments are
soon obliterated (fig. 23). Four or at most five cycles only can be recog-
nized. The sextants, however, may often be distinguished until quite a
late period (fig. 33).

The above-described order of the early divisions of the segment (w^alls
i to vii, figs. 14-22) is easily followed in its main outlines in good serial
sections of any leptosporangiate fern root. But English and German
text-books* still adhere unanimously to the statement of Nageli and
Leitgeb that the first division in the segment is the sextant wall, fol-
lowed by that which separates periblem and plerome. The first error was
corrected by Lachmann (1885; 1887; /ide Van Tieghem and Douliot), and

*Nageli and Leitgeb, 1865, 1868; Pteris Campbell, 1895, p. 328-329, fig. i6sA; 1905

hastata, plate 14, fig. 7. p. 333-

Sachs, 1875, pp. 124-125, fig. io2A. Strasburger, 1897, pp. 31 1-313, figs. 139, 140.

De Bary, 1884, pp. 18-19, fig-s- ?A, SA. Sadebeck, 1898, p. 61, fig. 4iA.

Goebel, 1887, pp. 214-215, fig. 162. Strasburger, Noll, etc., 1898, pp. 150-151,
Bower, 1889. fig. 165.

Vines, 1894, pp. 149-150, figs. 114, 115- Haberlandt, 1904, p. 74, fig. 14.

12

STRUCTURE AND LIFE-HISTORY OF HAY-SCENTED FERN.

the second by Van Teighcm and Donliot in 1888. The latter authorities
indicate that in some ferns (Pten's, Adiaiilum, Aneiinia, etc.) the outer
cell (fig's. 14, 16, 17) may give rise to two or three layers of cortex in ad-
dition to the epidermis (see table of cell-lineage in fern roots, p. 46). Such
ferns are in the minority. The same writers state that in Eqn isctian,
Osmunda, and Todca the first periclinal wall is between the central cyl-
inder and cortex, but that this is not the case in any other Pteridophytes
which have a single initial in the root.

I have found the account here given for Dcunstcedtia as to the origin of
epidermis, hypodermis, cortex, and endodermis to apply equally to root-tips
of Cibothun regale, Aspidiiun molle (fig. 48), Lygodium japonicum, OiiocJea
sensibilis (fig. 51), Ceratopteris thalictroides (fig. 50), and Asfiiciiuin mar-
ffina/c (fig. 47). In Pieridium aquiliuum and Didymochlccna lunulata (fig.
49) the epidermis and two layers of cortex are derived from the same part
of the segment.

Above the region of cell division in the root-tip there intervenes a brief
region of elongation. Beyond this, viz, about 2.5 mm. from the apex,

root-hairs appear. Each hair is

... , . ... ,. TAP.LE \. — Root-hairs.

a cylindrical outgrowth from the

lower (distal) end of an epidermal

cell. The cavities of cell and hair are

continuous, and contain but one

nucleus (fig. 252) lying variously

in the wall-layer of protoplasm.

In functional hairs the nucleus is

seen near the swollen apex. The

walls of the hairs are thin, of a clear, brownish-yellow color, and are often

molded around irregular particles of earth. Table 3 gives the exact

dimensions.

A transverse section of the region of functional root-hairs (figs. 34, 35,
44) shows the epidermis, hypodermis, four or five (rarely three) layers of
cortex and a well-defined endodermis. A single layer (rarely doubled in
places) of pericycle surrounds the cylindrical, diarch bundle. Protoxylcms
abut directly upon the pericycle at diametrically opposite points, and
between them lies a group of two to four (rarely five) large tracheids.
Extending around within the pericycle from each side of each protoxylem
is a row of three to seven sieve-tubes. Midway between the protoxylems
and lying against the pericycle is the small-celled, dense protophloem.
Between the phloem and xylem are cells of conjunctive parenchyma.

The epidermis (piliferoiis layer of Van Tieghem, etc.) at the level we
are speaking of is fully mature, and consists of cells four to six times as
long as wide. In cross-section they are nearly isodiametric, of slightly
variable depth and width, and bulging out a little on the outer side. The

Length.

Diameter.

Remarks.

mm.

mm.

0.5 +

0.014

Broken off.

0.3+ 0.014

Do.

O.2

0.014

Entire; immature?

SPOROPHYTE.

13

walls are brownish-yellow, like those of all of the cells outside of the
endodermis. The subjacent hypodermal cells are about three times as
wide and twice as deep radially as those of the epidermis, and similarly
elongated. The first cortical layer is composed of cells nearly twice as
large in cross-section as the foregoing", but like the last in length and
character of wall. Intercellular spaces occur rarely at the angles of these
cells. The cells of the second cortical layer are smaller again, often as
narrow as the epidermal cells. The two innermost layers are still smaller.
In the last three layers, especially the middle one, thickening of the walls
begins even before the root hairs are fully mature (fig. 43). At this stage
the endodermal cells are already very long, narrow, and practically empty.
Pericycle and conjunctive parenchyma are full of dense, granular contents.
They are probably multinucleate, since the cells are very narrow and long,
but nearly always show a nucleus in cross-sections. The pericycle cells are
often much larger in the region of the protophloem than elsewhere (fig.
35). The protophloems have already passed their greatest density and
prominence and the sieve-tubes now appear mature. Each protoxylem
consists of one or two extremely slender spiral elements, with one or two
slightly wider scalariform tracheids on either side. The two or four large
central tracheids of the metaxylem show as yet no thickening of the walls
(figs. 34, 35, 36, 43, 44).

TABLE 4. — Statistics of transverse section of root.

No. of cells in—

Cortex.

No. of cells in—

Epiderm.

Hypoderm.

No. of
layers.

No. of cells
in outer
layer.

No. of cells
in inner
layer.

Endoderm.

Pericycle.

64

30

4

*22

20

19

21

49

!9

4

M

24

18

20

40

i?

4

...

...

...

...

40

20

4

M

23

16

17

46

26

...

*i8

13

10

13

t6i

27

...

22

20

15

'9

I52

28

4 or 5

...

21

15

20

150

26

3

21

17

13

16

i with 3

14 with 4

10 with 5

3 with 6

*Immature. f Sections of one root, 16 p. apart.

Following all these parts upward in an old root, we find that the epider-
mis and outer soft layer of cortex wither after the root-hairs die, and are
ultimately sloughed off (fig. 46). The bundle is now protected by the
two or three inner cortical layers, whose walls have thickened so as almost
to obliterate the lumen. Lignification in the metaxylem takes place slowly.
In a section showing a withering epidermis, and the inner cortex indurated

14 STRUCTURE AND LIFE-HISTORY OF HAY-SCENTED FERN.

so that the cell -walls arc abcmt one-fourth the width of the lumens, there
are (fig'. 44) two large tracheids slightly thickened, and two in the center
wholly unthickened and containing protoplasm. Only with the decay and
disruption of the outer layers of the root do the inner cortex and xylem
become fully mature (fig. 45). This occurs from 3.5 cm. to 10 cm. from
the root tip. One can not as a rule, obtain a transverse section showing
all of the tissues mature and intact.

These facts have an important physiological significance. We suppose
the water-supply of the plant to come in through the root-hairs. But
where these are functional the xylem, always considered the water-con-
ducting tissue, is decidedly immature. It is evident that lignified walls are
not necessary for the conduction of water in the cells. It may be, however,
that the inner cortex, whose walls thicken at such an early period, is fora
time active in water conduction.

In 25 out of 95 roots examined (/. <?., 26.3 per cent) from localities in
Long Island, Pennsylvania, and Maryland, a more or less copious growth
of non-septate fungus hyphre was found in the middle cortex (figs. 52,
53). One of the most pronounced cases was attached to a rhizome of
unusual thickness and width and with an unusually long and rapidly grow-
ing meristematic apex. In a root-tip from this plant, hyphse were seen
3.07 mm. from the initial cell. They were located in the epidermis and
hypodermis, with branches running inward. In an older root, a hypha
was seen extending into the root through a root-hair (fig. 53). Once in
the middle cortex, strong hyphae run from cell to cell, ramifying in each
cell to form a dense granular mass (fig. 53). Sometimes the fungus is
found only on one side of the root, or in only a few cells. Occasionally
it spreads all the way round. From the undoubted vigor of the host where
the fungus occurs, the early stage of the root at which it appears, and the
mode of copious branching of the hypha^ in the mcdio-cortex, we feel
justified in considering that we have to do with a true mycorhiza. On
account of its inconstancy, it may be called facultative. No other poly-
podiaceous fern is known to possess such a commensalism, though Jansc
(1895) has recorded a similar condition in the aerial roots of Cyathca
(species not given).

Lateral rootlets arise from rhizogcnous cells of the endodermis, opposite
the xylem rays, as is universal in ferns. The endodermis has been shown
(p. 10) to originate in the root apex with two cells in each segment, /. <?.,
one in each sextant. After the segment is divided transversely into two
layers, there are two endodermal cells to a sextant, one lying nearer the
apex than the other. In the two opposite sextants in which the protoxy-
lems are later to form, the endodermal cell lying nearer the root apex
remains undivided by any radial walls, while its posterior and sister cell,
like all the other endodermal cells of the root, is halved radially. This
large cell, extending clear across the sextant, is the primitive rhizogenous

SPOROPHYTE. 15

evil (figs. 32, 33, 34). On one side it lies in a major sextant, on the other
side in a minor; but both of these sextants remain much narrower than
the other four (figs. 32, 33), as if to accommodate this undivided cell.
As the root elongates in this region, transverse divisions take place in
the rhizog'enous cell. The more distal member in each case retains its
identity, while the others sooner or later divide radially and become like
ordinary endodermal cells (figs. 38-40). If the radial walls are slow
to appear, we may have a row of three or four equally broad cells (figs.
38-40), of which only the most distal is rhizogenous. In Cyatheacea.- all
of the cells in the same vertical line with the rhizogenous cell are said to be
of the same width as the latter (c. g., Cibolhii/i regale; cf. Sadebeck 1898,
p. 63). Dennst&dtia, therefore, agrees with Polypodiaceae in this respect.

After elongation of the root is complete, the definitive rhizogenous cell
swells out into a lens-shaped body (fig. 34). On its proximal side it is
cut once more by a transverse wall, but this wall passes obliquely inward
(fig. 37 ) . Two obliquely longitudinal walls follow (figs. 35, 36, 40, 41) hew-
ing out a tetrahedral cell with one face against the cortex, one side toward
the stem, and the apex toward the subjacent pericycle. This is the rootlet
initial (figs. 36, 40, 41), and the three cells which bound its sides are the
first segments of the rootlets. This initial and all of its segments proceed
to develop as in the case of the parent root (figs. 41-43). The rootlet
being smaller than the root (fig. 54), its segments undergo fewer divisions
in the cortical region. The xylem of the rootlet stands transversely to
that of the root. Therefore the protoxylems are to the right and left if the
root is held in a vertical position, and the xylem band of the rootlet will
lie horizontally. The xylems of the rootlet arise in the second and third
(and overlying) segments formed from the rootlet-initial, and in their
proximal sextants.

Meanwhile divisions have also occurred (as shown by mitotic figures)
in the neighboring endodermal and cortical cells. In the former no regu-
larity was observed. The cortex, however, develops a mass of small pro-
toplasmic cells directly over the rootlet (figs. 34, 42, 43) and undergoes
no induration here. The cells immediately adjacent to the root-cap are
finally organized into a special layer (fig. 43) which advances through the
remaining cortex, etc., apparently causing the disintegration and absorp-
tion of these tissues. In the mature stage all of the tissues of root and
rootlet, excepting epidermis and outer cortex (including hypodermis), are
respectively continuous. Certain inner cortical cells of the main root bend
out into the branch, but eiidodermis and pericycle (fig. 55) connect by the
intervention of a number of cubical cells. The xylem tracheids of the
branch terminate abruptly against the side of those of the main root and
at right angles to them (figs. 56, 57). The phloems connect in a manner
similar to the xylems. The mature rootlet repeats the structure of an
ordinary root on a smaller scale (fig. 54). In a slender rootlet there may

16 STRUCTURE AND LIFE-HISTORY OF HAY-SCENTED FERN.

be but two layers of true cortex between endodermis and hypodermis . In
this case the outer layer has six large cells in a circle, and the inner layer
and hypodermis have twelve cells each. The sextant walls of the apex
thus persist, and may even at times be traced through the endodermis and
pericycle. Plainly this does not in any sense indicate a common origin of
endodermis and pericycle.

In relation to the stem, the main root originates very near the apex, in that
layer of cells which will subsequently give rise to both endodermis and peri-
cycle (fig. 70). The rhizogenous cell is large and cubical. By three oblique
walls a tetrahedral initial is early cut out, and from this time onward it
behaves just as it would in a mature root. It has no fixed position in rela-
tion to the stem-initial. Beneath (centrad) the developing root-tissues a
great proliferation of stem-tissue takes place. The root is thus borne half-
way through the cortex, or farther, on a "pedicel" (figs. 59-61) while it
is of itself digesting away the outer cortex and forming a many-layered
cap. It finally emerges through a ragged opening in the cauline cortex
(fig. 102).

During the elongation of the stem the vascular tissues connecting root
and stem are much deformed, with the following result: Tracing the
mature root inward, its stele passes obliquely half-way through the cortex
of the stem (fig. 102), then bends sharply backward, and fuses with the
stem stele. The cortex, endodermis, and pericycle of both organs are
smoothly continuous. But the root-xylem, after passing a little backward,
turns at an acute angle forward for a short distance. Some elements
become at once cauline, and run on forward. Others, by another sharp
bend, turn backward in joining the xylem of the stem. Rarely a tracheid
runs directly from the root into the stem without a double bend. Phloem
and conjunctive parenchyma follow parallel with the xylem. The depth
of these bends varies. In any case, those tracheids which are continuous
from root to stem must assume very peculiar shapes (figs. 58, 61, 62).
The bending occurs at an early stage of the development, before the cells
become lignified (fig. 61). When a root originates from a leaf-base it
passes out in a similar manner. A double bend occurs in the vascular
elements, but the folds are quite shallow.

THE STEM.

The rhizome is slender and cylindrical and more or less branched (fig.
3). A piece in my collection from Bucks County, Pennsylvania, grown in
loose loam, is 35 cm. long, with branches 15 cm. and 4 cm. long. Another
rhizome from Delaware County, Pennsylvania, is 8 cm. long to a fork; one
branch runs 23.5 cm. and forks into parts 4.5 cm. and 3 cm. long, respec-
tively; the other branch runs 19.5 cm. to a fork, with the parts 22 cm. and
13 cm. long. The diameter varies from 1.5 mm. to 4 mm., with a mean
of about 3 mm. When fresh the rhizomes are somewhat flexible, but the

SPOROPHYTE. 17

cortex readily breaks across and pulls loose from the much more tough and
elastic vascular core. The outer surface is of a dark reddish-brown color,
shading' to nearly white at the apices. In very rapidly growing- rhizomes
the soft white apical region may be 2 cm. long-, but usually it is about
5 mm. Glabrous to the unaided eye, the mature parts of the rhizome may
be seen with a hand lens to be scattered over with the bases of dead hairs.
At the apex the hairs themselves are very numerous and serve to clothe
the soft parts with an efficient protective covering- (fig's. 70, 106). Scales
are entirely lacking- from all parts of the plant.

The leaves are inserted without articulation, mostly alternately to rig-ht
and left of the mid-line on the dorsal surface of the rhizome. They stand
therefore in two dorso-lateral rows.* Since the distances between them
rang'e from 1.6 cm. to 5.7 cm. (usually about 3 cm.), we may speak of
definite nodes and internodes (fig-. 3) (</. Boodle, and Gwynne-Vaug'han).

Branching" occurs in two ways: direct forking- of the rhizome, and occa-
sional stem-buds given off from the lower parts of the petioles (figs. 3,4).
The latter will be discussed in connection with the leaf. A fork gives all
the appearance of a true dichotomy, the two branches spread equally, like
the arms of a Y, from the parent axis, and a ridg'e runs over the crotch
(fig-. 99) from dorsal to ventral surface of the stem. In young- stages
(arms 1 to 2 cm. long") the two branches are alike in length and diameter,
but inequalities of growth soon cause them to differ. I did not determine
how the two initial cells originate.

The interesting and beautiful concentric structure of the stem (fig. 67)
was evidently known to De Bary (1877, 1884), and was thus described by
Eaton (1879):

The section shows a broad exterior ring of light-brown parenchyma; inside of this
is a broad circle of minute white starch-cells, then scalariform vessels in a narrow ring,
bordered by other minute cells, which are most probably bast cells; inside of this is an-
other broad circle of starch-cells and in the very center is a roundish mass of brown
sclerenchyma. The whole section has such a regular concentric system that it is not
only very pretty to look at, but would be very well suited for anatomical study in the
classroom [p. 341].

Gwynne-Vaughan (1903) also says:

A perfectly solenostelic vascular system was found in the stems of all the species
included in the following list: Davallia hirsuta, marginal:*, strigosa, platyphylla,
hirta, spelunccc, novce-zeylandia, Lindsay a retusa, Dicksonia apiifolia, ctcutaria, scabra,
punctiloba, davallioides, Pteris scaberula, incisa,ludens, Pellcea atropurpurea,falcata,
fatnesonia imbricata. All these ferns have a creeping, more or less dorsiventral rhizome,
with the leaves arranged in two rows on the upper surface, and their solenosteles differ
from each other and from that of Loxosoma as described in Part I of this paper [1901]
in so slight a degree that the same description will serve for them all [p. 691].

*Of 134 leaf-bases on stems which were taken at random, 96 insertions were dorso-
lateral (right or left), 21 directly on one side (right or left), 16 dorsal, and 3 ventral.
Each ventral leaf springs from the angle of a fork of the stem, and bends upward
through the fork.

18 STRUCTURE AND LIFE-HISTORY OF HAY-SCENTED FERN.

In the mature internocle (fig's. 63, 64, 67, 97) we find externally an
irreg'ular blaek, sclerosed epidermis (see table of measurements p. 19).
The inner walls often bear rod-like thickenings projecting into the lumen
of the underlying cells (fig. 63). Next within comes a broad zone of black-
walled sclerotic cortex, ten to twelve cells dee]). These cells (fig. 63, s)
are angular and do not inclose intercellular spaces. Although mostly
appearing empty, they may contain considerable amounts of starch, espe-
cially in the inner members. They are largest in the middle of the zone.
There are pores of different sizes and shapes irregularly scattered over the
walls. The ends of the cells are oblique or bluntly tapering (fig. 64; see
table, p. 19).

Sharply demarcated from the sclerotic cortex by a sudden change in the
character of the wall is a narrow layer of starchy cortex (figs. 63, 64, s/.).
This layer is from three to eight cells thick. The cell- walls are thin and
colorless. They inclose protoplasm and nucleus, and are densely packed
with starch. Intercellular spaces are easily perceptible here between the
rounded cells. The end walls are transverse or are slightly oblique (see
table, p. 19).

The cortex is bounded on its inner side by a continuous sheath of endo-
dermis — a circle, in cross-section, of flattened but irregular cells, very poor
in contents (figs. 63, 64, 67, 97). Each radial wall bears a thickened band
which stains with safranin (see table, p. 19).

Within this layer intercellular spaces can be no longer found, but we
encounter a thick ring of vascular tissue, which again incloses a rod-like
core of starchy and sclerotic cells (medulla). In short, the vascular sys-
tem constitutes a cylindrical tube. The vascular ring itself shows, passing",
from without inward, successive rings of-

(1) Endodermis. (7) Inner conjunctive parenchyma.

(2) Pericycle. (8) Inner phloem.

(3) Protophloem. (9) Inner protophloem.

(4) Phloem. (10) Inner pericycle.

(5) Conjunctive parenchyma. (n) Inner endodermis.

(6) Xylem.

In the words of Gwynne-Vaughan (1903):

The xylem ring is surrounded both externally and internally by a complete ring of
phloem and pericycle, and the whole is delimited from the ground tissue on both sides
by a well-marked endodermis.

The outer perieycle (fig. 63) consists of two or three layers of angular
evils, densely packed with starch. Their walls tend to lie in radial and
tangential planes, and the radial walls are often in continuity with those
of the endodermis (see table, p. 19).

Protophloem is difficultly discernible in the old stem, though it forms an
almost continuous ring of cells. The elements are greatly elongated, very
slender, angular, of various shapes and sizes, with finely tapering ends.

SPOROPHYTE.

19

Only rarely do two elements lie adjacent radially. The walls are thick,
but not ligirified, and are beset with deep pits with rounded, flaring aper-
tures. They are best seen in the very young regions (figs. 107, 110)
where lignification is just beginning to occur in the xylem (see table bel< >\v ) .
The phloem forms a continuous ring, one to three cells in thickness.
The cells are ang'ular and appear in cross-section to be empty; their ends
(fig-. 65) acute, but not finely tapering-. These are sieve-tubes, but the
sieve-plates are only found with some difficulty. I could not demonstrate
callus, either with azoblue or coralline soda.

TABLE 5. — Measurements in millimeters of cells of various tissues in the stein.

Tissue

Length.

Radial diameter.

Tangential diameter.

A\vr.

Max.

Min.

Aver.

Max.

Min.

Aver.

Max.

Min.

Epidermis

O. I

.264

•336
.142
.063
.01447
.678

o. 16

0.064

O.O2
•037

.051
.042
.00825
.0157
.00892
.0164

OI ^7

0.03
.047

.066
.05
.009
.0178

0.014
.032

.032
.032
.00714
.014

0.035

0.0372

0.032

Outer sclerotic cor-
tex

Inner sclerotic cor-
tex

•44
.246
.078
.16

. 21

.088
.057
.127

Starchy cortex

Outer endodermis...
Outer pericycle

.0195
.0239

.0214
.032

.014
.0178

Outer protophloem..

2. I

i-5

.025

.013

Outer conjunctive
parenchyma

-VT- 1

. 167

4-4

1.6

• wl -/
.06

•0155
.0083

•034

.078

-O2I
.OI4
•039

.032
.0134
.0071
.025

.04

.057

.02

Inner endodermis...
Starchy medulla

.099
.166

.149
.264

.082
.105

.01985

.022

.0178

Between phloem and xylem lie the thin-walled, angular, starch-laden cells
of conjunctive parenchyma (figs. 63, 64, 81). They are about twelve
times as long as wide, and have tran verse or oblique end- walls. Some of
these cells may extend in between the xylem tracheids, and rarely such an
extension joins with the inner conjunctive parenchyma, completely inter-
rupting the xylem. Sometimes the conjunctive parenchyma is interrupted
and a sieve-tube lies directly against a scalariform tracheid.

With the above exception the xylem is a continuous ring of larger and
smaller scalariform tracheids, with thick, lig-nified walls (figs. 63, 67, 97).
The middle lamella is clearly discernible. In places this ring includes but
one radially widened xylem element; in other places it may be three cells
thick. There are no spiral elements in the stem. The tracheids are of
the usual long-, pointed type (figs. 78, 79; see table above).

The inner tissues so closely resemble the outer that they may be described
simply by comparison (figs. 63, 64). Conjunctive parenchymas show no
difference. The inner phloem is mostly one, often two, cells thick, and the
sieve-tubes are narrower than the outer ones (see table above). Proto-

20 STRUCTURE AND LIFE-HISTORY OF HAY-SCENTED FERN.

phloems present no differences. The inner pericyele is one to two, rarely
three, cells thick. Inner endodermis is like the outer in cross-sections, but
seems to have longer cells (see table, p. 19).

The layer of starchy medulla inside the vascular ring- corresponds with
that on the outside. In a rhizome with six to eight layers outside, there are
about seven inside. In another with three or four (mostly three) layers
outside, there are two or three (mostly three) inside. But the inner cells
are smaller than the outer (see table, p. 19). The inner sclerenchyma
cells (sclerotic medulla) are longer and narrower than the outer, and have
thicker and blacker walls (fig's. 63, 64, 76, 77, 80). They form a core
from twelve to twenty-two cells in diameter.

The above type of stem is called by Gwynne-Yaughan (1901) asolowstc/c
(adj. solenostelic) and by Jeffrey (1897) an am[>hif>hloic siphonostclc. The
description may serve as a definition of these terms.

At the node (fig's. 3, 4, 66) the cylindrical leaf-base spring's from the
slig'htly larger stem at a right or acute angle, usually without altering- the
size or shape of the stem. Occasionally the stem is slightly enlarged
below the node, and rarely there is a slightly prominent ridge between leaf
and stem, as at a fork.

The leaf-trace or vascular bundle of the leaf (petiolar meristele) leaves
the stem as a trough-like band (horseshoe-shaped in transverse section)
which is of the same thickness as the wall of the vascular tube of the stem
(fig. 82). The concavity of the trough faces obliquely upward and forward
in most leaves. But where the leaf-insertion is ventral (figs. 83-87) or
lateral the trace faces directly upward. When the insertion is dorsal the
trace faces directly forward. At the place where the leaf-trace leaves the
tubular vascular system of the stem a distinct leaf-gap occurs (fig. 82).
This is a narrow slit in the stem bundle, through which the medullary and
cortical tissues become continuous. The gaps differ in shape and in their
exact relation to the leaf-trace. One gap is 11 mm. long and 1.2 mm.
wide, with acute ends, and with the leaf-trace attached near the middle of
the ventral side. Another is 14 mm. by 1 mm. A third is 1.8 mm. by
0.3 mm., rounded at both ends, with the leaf-trace occupying- nearly all of
one side. The average size (of ten) is 5.45 mm. long and 0.53 mm. wide.
I'sually the anterior end is rounded and the posterior end tapering and
acute, with the leaf-trace attached along one side, at or near the anterior
end. Such a gap is figured by Gwynne-Yaughan (1903, plate 33, fig. l).
When the leaf-insertion is dorsal the trace arises symmetrically from the
rounded posterior end of the gap. Lateral leaf-traces differ scarcely at all
in their origin from the usual dorso-lateral type. Ventral leaves, spring-
ing from a fork, are symmetrically attached to each arm of the stem (figs.
83-87) . The trace lies like a trough with its concavity upward. Approach-
ing- such a nodal fork along the main stem, the stele first becomes wide
and flat. Then a slit is found on the upper side, and as the branches

SPOROPHYTE. 21

separate this slit forms a leaf-trap in the adjacent (inner) face of each
branch. From the lower margin of both gups the trough -like leaf-trace
comes off, forming- for a time a connection across the fork from one branch
to the other. It soon becomes distinct from both. Leaf-gaps are easily
dissected out, since the tissues readily separate along the line of the endo-
dermis.

As stated above, the cortex and medulla of the stem come into contact
or continuity through the leaf-gap (fig. 100). The starchy layers always
connect, and sometimes a strand of sclerotic cells connects the outer cortex
with the similar central medulla. In about half the nodes examined (9 out
of 20) the sclerotic medulla passes the leaf -gap as a solid rod unchanged.
In one-fifth (4) it connects directly with the outer cortex through the
wide leaf-g'aps. In 7 a rod of sclerenchyma passes from the medulla out-
ward in the groove of the leaf-trace to vanish in the petiole or to become
continuous there with a peripheral layer of similar cells. This rod may
originate independently in the starchy medulla shortly below the node, or
it may begin as a ridge on the sclerotic medulla, which is gradually con-
stricted off. These different arrangements of sclerenchyma may occur at
successive nodes of one stem.

Between the cortex and the xylem of the stem, all of the inner and outer
tissues become continuous around the margins of the leaf- gap— endodermis
with endodermis, pericycle with pericycle, phloem with phloem, and con-
junctive parenchyma with conjunctive parenchyma (fig-. 108). Or, the
phloem may become very thin, or may be completely interrupted for a
short distance on one or other side of the gap. A transverse section of
the stem through a leaf -gap shows the vascular system as a deep, round
horseshoe (fig. 100). No new tissue elements are seen at the nodes.
Spiral vessels are not found in leaf-trace or stem. The cells bend out
from stem to leaf by g'entle curves, without any noticeable peculiarities.

Where the stem forks each tissue system remains continuous and
unbroken (fig. 99). There is no ramular g'ap. One can best imagine the
structure by starting with a Y-shaped object made of round rods, welded
together below. Let this represent the sclerotic medulla. We dip the
object into melted wax, coating it all over; this represents the layer of
starchy medulla. Successive coatings of suitable thickness may represent
inner endodermis, inner pericycle, inner phloem, xylem, outer phloem, outer
pericycle, outer endodermis, starchy cortex, sclerotic cortex, and epidermis.
In the angle of the Y the continuity of tissues from one arm to the other
is strikingly smooth and regular. A single scalariform tracheid may
extend for a long distance in each arm. On the sides of the fork this con-
tinuity involves angular elements of peculiar shapes (fig. 98).

The apex of the stem is so clothed with hairs as to appear smoothly
rounded (fig. 3). Under this covering there is a shallow, basin-like de-
pression (fig. 70), not often symmetrical, at the tip of the stem. A low

STRUCTURE AND LIFE-HISTORY OF HAY-SCENTED FERN.

protuberance standing out in the center of this basin is the growing point.
Its surface is naked and is rendered irregular by leaf -rudiments. At the
top of the protuberance is the stem-initial— a narrow, irregularly triangular
cell. In longitudinal section it is deeper (0.0642 mm. to 0.07 mm.) than
any other cell, excepting a newly formed segment (figs. 106, 129). But
in cross-section it is smaller than many of its near neighbors (figs. 71, 103,
109). This narrowness makes it difficult to recognize. Further, as the
protuberance in which it stands varies in apparent position according to
the development of the young leaves, one can not be sure of getting satis-
factory sections of it by cutting a bit of stem in an exactly transverse or
longitudinal plane.

Segments are cut off in regular succession on the three interior faces of
the initial. As I could not, even with much effort, distinguish any regular
position of the initial with regard to dorsal and ventral surfaces of the
stem, we can not speak of dorsal or ventral segments. The irregular
arrangement of the leaves is probably related to this irregularity of the
position of the stem-initial. The order in which the segments are cut off
is cither from left to right or right to left, the two occurring in about equal
numbers in my preparations (7 counter-clockwise, 5 clockwise, from older
to younger).

Each segment is divided first by a periclinal wall near its inner end.
The small, deep-lying cell thus formed gives rise to the medulla (fig. 106).
The outer cell is next cut by the sextant wall — a radial anticline, dividing
the cell nearly into halves ("sextants"). This wall remains prominent
for a comparatively long time (figs. 71-75, 103, 109) G;/~. Bower's figures,
1889). The relative sequence of the following anticlines and periclines
was not determined. The second periclinal wall is formed near the inner
end of the columnar partial segment or sextant, giving an inner nearly
cubical cell and an outer columnar one (fig. 106). The inner of these
(plerome rudiment) gives rise to the vascular system of the stem, including
outer and inner endodermis. The sextants are usually halved longitudi-
nally at right angles to the sextant wall (fig. 73), and as many as sixteen
to twenty rectangular cells may be formed by walls at right angles to the
two just mentioned (figs. 75, 103). Very often oblique walls break up the
symmetry. The outer cells are repeatedly divided by periclines near the
inner end (fig. 106) until, after four or five such partitions, the remaining
< niter portion is reduced to the depth of the epidermis. Division then ceases.
Meanwhile the whole segment has been pushed farther and farther from the
growing1 point of the stem. The last pericline occurs in cells which are
nearly half-way tip on the sides of the basin-like depression of the stem apex.

Near the lowest part of this depression arise the hairs which clothe the
apex (fig. 106). A superficial cell bulges out slightly and is cut obliquely.
The outer member enlarges in length and diameter, and is divided by
several septa. These divisions are often intercalary. The two basal cells

SPOROPHYTE. 23

also straighten up their originally oblique wall, until it stands perpendic-
ular to the surface of the stem. Thus there results a basal cell of a hair,
with a hairless sister-cell beside it. A similar development of protective
ramentum was described by the writer on the stem-tip of Nympha_ja.

Returning to the plerome rudiment, it develops much more slowly than
the cortex. It divides periclinally into two equal parts (fig-. 106) and cadi
of these again by similar walls, giving four layers of cells in the plerome.
Apparently the outermost and innermost of these .give rise to pericycle and
endodermis, while the two median probably produce xylem and phloem.
Certain it is that all the tissues just named come from the four cells in
question. It is also certain that the endodermis is formed at the last peri-
clinical division in the outermost layer of plerome, and each endodermal
element is a sister cell to the pericyclic cell radially next to it. This has
been suspected by several writers for several ferns, by reason of the con-
tinuity of the radial walls. I was able to prove it for Dennstccdtia. by find-
ing' a number of mitoses (</. fig. 70, 6).

The stem very quickly attains its final diameter; hence its broadly
rounded end. At a distance of 0.24 mm. from the initial cell, in an aver-
age instance, all the cells have reached their ultimate width. Already at
0.18 mm. in this case the epidermal cell-walls are thickening and becoming
dark brown. At 0.345 mm. the protophloem is mature and at 0.5 mm.
lignification is beginning- in the xylem. In another case, however, the
first thickening-s of xylem occurred at 6mm. from the apex, and in another
at 0.18 mm. (</. fig'. 70).

Excepting' the epidermis, the first tissue to mature is the protophloem
(fig'. 107). For some time it forms a prominent ring- of small, angular,
thick- walled fibers near each boundary of the vascular bundle. The walls
stain deeply with haematoxylin. Seen on the side, the cells are very long
and slender, and the walls are peculiarly and irregularly pitted. The xylem
matures very irregularly. A cell here and another there become lignified,
apparently without any order (fig-. 110). The larger tracheids precede
the smaller, the first ones being often near the exit of a root or leaf. No
spiral cells are formed, and there is nothing that could be called protoxylem.

THE LEAF.

In the taxonomic literature of ferns the chief attention is directed to the
leaf and its important appendages. Thus several references have already-
been made to the leaf of Dcnnst&dtia (figs. 1,2). In the following descrip-
tion we shall designate as petiole, the leaf-stalk, from the rhizome to the
lowest pinna. All above this is blade. The continuation of the petiole in
the blade we shall call rachis, and the. thin expansion of the leaf lamina.

All of the leaves are alike in general appearance and in structure. There
is no morphological distinction of sporophyll and trophophyll, although for
various undetei'inined reasons some leaves are fertile and others sterile.
Both kinds are produced throughout the season, though the greatest devel-

24

STRUCTURE AND LIFE-HISTORY OF HAY-SCENTED FERN.

opment occurs in May and June. Two to five leaves is the common annual
growth from each branch of the stem . With considerable accuracy each
leaf faces the chief source of light. When growing around an old stump
or a bowlder, the leaves on every side turn their backs to it. On steep
hillsides, also, the leaves face downhill. While not every leaf, by any
means, stands thus, the general effect is very noticeable. The leaves stand
nearly erect, but the blade is gently curved backward. A leaf which ma-
tures in June (first crop) turns yellow in August or September in eastern
Pennsylvania and Maryland, by reason of old age (whatever that may be).
In New England the first leaves persist throughout the season. The first
sharp frost kills all of the foliage of Dcnnsia-dtia.

TABLE 6. — Data of the leaf.

No.

Total
length.

Length of—

Maxi-
mum
width of
leaf.

Distance ol
widest part
from base
of lamina.

Diameter
of petiole.

No. of
pairs of
pinnae.

Maxi-
mum No.
of pairs of
pinnules.

Petiole.

Blade.

rm.

em.

cm.

cm.

on.

mm.

I

1 1 1

23

88

21

24

2-5

46 24

2

53 J3

40

l6-3

15

2. 2

37 22

3

33-5 7

26.5 6.8

12

1.8

3°

13

4

41 io

31 11.7 12.5 1.8 32

19

Av.of 10

60 . 38 14.2

46.18

16.49 15-7

2. O2

37-5

21.5

The petiole is slender, smooth, "chestnut-brown" (Clute, 1901), chan-
neled on the upper (ventral) surface, about one-fourth of the total length
of the leaf (not one-half, Clute, 1901). It is slightly stouter at base, e.g.,
3 mm. in diameter below and 2.5 mm. above. The lower part is nearly
always curved, often very much so, because of the obstructions the leaf
meets in coming out of the soil. Above the earth it is straight and nearly
erect .

The brown color of the petiole is due to pigment (phlobaphene ?) in the
cell- walls of the epidermis and outer cortex. At the base, the cortex may
be colored to a depth of nine cells. Passing upward along a newly ma-
tured leaf, the layer grows thinner, and just below the first pinna; there is
often no pigment at all. But the coloring advances upward with the age
of the frond, until all of the rachis may become brown. The epidermal
cells are very irregular in size. Near the rhizome they are polygonal and
are about twice as long as broad. They become longer above. At 1 cm.
from the rhizome they are five times, and just below the blade seven to
eleven times, longer than wide. In transverse section (fig. 69) they are
nearly is< (diametric, but of very different depths. Each cell bulges out a
little on the surface of the petiole. The cell- walls are thickly lignified,
and bear numerous pits suggesting a continuity of protoplasm from cell to
cell. There are many hairs, both glandular (fig. 249) and acicular (fig.
258, b~) on the young petiole, but they wither or are completely shed at

SPOROPHYTK. 25

maturity, leaving only the oval basal cell to show where they stood. The
hairs are of the same character as those on the lamina. They will be
described later.

Beneath the epidermis is a layer of brown, thick-walled sclerenchyma,
as already mentioned (</. fig'. 69). In the petiole proper these cells are
long-, and tapering at each end. Near the node they become shorter until
they reach the dimensions of the cells of the sclerotic cortex of the stem,
with which they are continuous. They may also unite here with a rod of
similar tissue derived from the medulla of the stem and lying- in the trough
of the leaf-trace ((/. p. 21). About the middle of each side of the petiole
the sclerenchyma is interrupted by an extension of ordinary parenchyma
out to the epidermis (figs. 69, 101). Stomata occur along this line to
admit interchange of gases between the intercellular spaces of the paren-
chyma and the outside air. For a short distance near the surface of the
ground the sclerenchyma is not so interrupted and stomata do not occur;
but at about 1 em. from the rhizome the stomata and parenchyma begin
again. The area occupied by them here is easily seen, because the walls
of the sclerenchyma cells are of a deep-brown color, while those of the par-
enchyma are colorless. This area may end 0.5 to 2 mm. above the rhizome,
or may continue as a visible ridge on the posterior side of the leaf-base
for 1 mm. or more along the surface of the node. The guard-cells of
these stomata stand above the level of the surrounding' epidermis. Such
lines of parenchyma are common in fern petioles, but by no means universal
(Thomae, 1886).

From the epidermis inward, the cells become larger in diameter, shorter,
less acute at the ends, and thinner walled. Two to ten cells in, according'
to height and age of section, we find soft parenchyma, with large watery
cells bearing chlorophyll. These become smaller again next to the vas-
cular bundle. They are in contact with the endodermis all around the
bundle. This layer has many intercellular spaces in its middle portion.

The bundle itself is shaped like a trough, opening- toward the upper
surface of the leaf. In transverse section near the rhizome (fig's. 82, 100, 101)
it is U-shaped, or very nearly like the periphery of a semicircle. Its thick-
ness is about the same all round. Higher up the petiole it is more V-shaped,
flattened at the angle (fig. 68), and swollen at the tips of the short arms.
The endodermis bears the typical Caspary's band. The contents of its
cells take up stains with avidity, and probably include mucilage. A par-
enchymatous layer within the endodermis may be termed pericycle. Of
two cells thickness in most places, it thins off to one cell around the tips of
the arms of the bundle, and becomes three-celled on the outer side of each
arm (fig'. 68). At the latter place, also, the middle cells enlarge greatly to
form a very peculiar open tissue. The thin band of xylem in the midst of
the bundle follows the latter in g'eneral shape, but is more curved in its
outlines. At the ends of the U or V it approaches very near to the peri-

26 STRUCTURE AND LIFE-HISTORY OF HAY-SCENTED FERN.

cycle and the tips are bent sharply inward or ' 'hooked' ' (figs. 68, 69, 88-92) .
The "hook" is short at the base of the petiole, and much deeper above.
Around the xylem there is a layer of conjunctive parenchyma. Between
this and the pericycle lies the phloem. Its outermost layer all round is
early developed as protophloem like that of the stem. It appears after the
protoxylems, and develops first on the outer side of each arm of the bundle.
vSoon afterward it appears on the inner side of each arm. These four patches
spread by addition of elements at each end until they meet and form a
continuous sheath. The bulk of the phloem consists of sieve-tubes. No
protoxylem elements (spiral cells) are found at the base of the petiole, but
they appear shortly above, and are on the inner (tipper, ventral) side of
the xylem. The first to appear are two patches of slender spiral, annular,
and spiro-anrmlar elements, one in each corner of the flattened angle of the
V-shaped xylem. Later another group (and finally two) appears on each
arm beneath the hooked end of the xylem. Thus there are six protoxylems
in all (fig-. 68). They develop when the petiole is very young-. But a
very great enlargement of the tissues, both in length and breadth, takes
place later, by which the protoxylem elements are quite torn to pieces.
Neighboring- parenchyma cells then push in to occupy the space, and the
remaining lignin rings and spirals are often pushed out of their original
places (figs. 116, 117). Thus originates the "cavity-parenchyma" of
Russow (1871) (</. Gwynne-Vaughan, 1901, p. 87). Its cells retain a
fairly regular rectangular shape and are thin-walled.

The blade is from 30 to 90 cm. long (see table 6) and 7 to 20 cm. wide,
lanceolate in outline, with the widest part from 5 to 40 cm. above the base.
It is pubescent, very soft in texture, and quickly wilts when plucked or
frosted. Its delicate yellowish-green color has been mentioned above. It
is thrice pinnate (fig. 2) with the margins of the ultimate segments deeply
crenate (fig. 5). The veins are forked, without anastomoses. The verna-
tion is of the typical circinate kind out to the divisions of the third degree.
The crenations develop as the leaf unfolds.

The number of pinnae varies from 30 to 50 pairs, and the maximum
number of pinnules on a single pinna is 13 to 25 (see table 6). Thoug'h
we speak of pairs of pinnae, they are not exactly opposite. One of the
lowest pair may be from 1 to 7 mm. below the other. Those of the next
pair, 2.5 to 8 cm. above, are similarly separated from one another. Farther
up, the pairs of pinnae and the individuals of each pair are closer together,
but the separation of the members of each pair is sufficient to make them
appear almost alternate. Near the apex of the leaf the rachis is much like
a midrib, and the divisions of the leaf diminish in size and degree. The
uppermost "pinnae" are mere crenations of the margin of a winged rachis.
But morphologically they are pinnae. The larger pinnae, in a similar
manner, have large, twice -divided pinnules; but these become smaller the
farther they are from the rachis, until they, too, become mere crenations.

SPOROPHYTE. 27

The rachis of the leaf repeats in general the structure of the petiole.
It is deeply channeled above, rounded beneath (fig. 69). The epidermis
is very thick-walled, and is supported by a layer of sclerenchyma. This
layer is thinner than it is in the petiole, especially on the back and sides of
the rachis. In the ridge on either side of the groove it remains strong.
About midway of each side of the rachis there is a line of spongy paren-
chyma with chloroplasts. Here the sclerotic sheath is interrupted and
stomata are found in the epidermis (figs. 69, 114). The vascular bundle
becomes narrower as we go up the leaf. From a very short, thick, V-shaped
cross section it finally becomes oval. Now, the xylem is a band with
hooked ends. The protoxylems are only two, lying inside the hooks.
Cavity parenchyma still accompanies them (fig. 114).

The oval form of bundle just described is also found in the base of the
rib of the larger pinna.'. These ribs, indeed, fully repeat the outlines and
structure of the upper part of the rachis. The vascular bundle of the rib
springs from that of the rachis in the following manner (</. figs. 88-92).
The hook of the xylem at one end of the U-shaped petiolar bundle grad-
ually becomes twice as deep as before, and a bridge of xylem is formed
across the middle of the hook . A constriction now cuts through the bridge ,
separating from the bend of the hook a small ring of xylem filled with
phloem. The constriction later affects the endodermis, and the new bundle
is completely separated. As it bends out into the pinna, the xylem ring-
grows thicker below and thins out to nothing above, until only a single
transverse band of tracheids remains in the middle of the bundle. A strand
of protoxylem lies on the upper side of this band, and is continuous with
that of the rachis.

The veins and veinlets of the leaf are collateral in structure, cylindrical,
with distinct endodermis. Their xylem consists wholly of spiral tracheids.
The ribs of the uppermost pinna? are also collateral throughout, with xylem
above and phloem beneath.

The structure of the lamina is very simple. An unbroken epidermis of
wavy-margined cells (fig. 112) covers the upper surface. If any cuticle
is present, it is extremely thin. Over the veins the cells are elongated and
approximately rectangular. The lower epidermis (fig. Ill) bears many
large oval stomata about 0.023 mm. by 0.038 mm. There are about eig-ht
of them per square millimeter. Thej^ are set exactly level with the
epidermal cells. Copeland speaks of them as follows (1902, pp. 349, 350):

In the Polypodiacens I have examined, there is an approach to what Haberlandt
calls the type [of stoma] of swimming plants, in that the ridge of entrance is well
developed, while the ridge of exit is inconspicuous or not present. In Dennsta'dtia
punctilobula Bernh. (figs. 40, 41) \cf. our fig. 104] this thickening of the ridge of entrance
has gone far enough to give the stoma a rigid appearance, but it is really motile.
Opening seems to be effected by a movement of the ridge of entrance outward as well
as backward, such as must occur in lesser degree in the case of Angiopteris. The guard-
cells of Dennsttctltia are thin-walled and shallow at the ends.

28

STRUCTURE AND LIFE-HISTORY OF HAY-SCENTED FERN.

In the following table the depth is measured from the ridge of entrance down to
the deepest dorsal focus. The stoma was closed by displacing water with alcohol.

Open.

Closed.

Width of stoma

M

T.O

M

28 5

Width of guard-cells

J^

13. ; to 14

J • J
14 tO 14 . S

Width of pore..

2. s

o

Length of stoma

44

4C

Depth of stoma

17

14

[P- 35°] The first effect of the alcohol is to widen the pore, which then gradually
closes, the sides becoming apparently straight before they meet. The increase in depth
at the ends, which is partly responsible for opening the pore of this stoma, works to
better advantage than in the stoma of Osjintnda.

The cells of the lower epidermis (fig. Ill) are smaller and more wavy
than those of the upper surface. All of the epidermal cells contain chloro-
plasts, but they are especially abundant in the gtiard-cells. The thickness
of the lamina runs from 0.06 mm. at the veinlets to 0.09 mm. where the
mesophyll is well developed. The air-spaces are especially large in the
lower half of the spongy parenchyma. A continuous layer of irregular
cells lines the upper epidermis, but there is nothing that could be called
palisade tissue (fig'- 113).

Scattered plentifully all over the leaf are hairs of two kinds — acicular
(fig. 258, b*) and glandular (furs. 105, 113, 249). The first are simple, acute,
septate, often 1 to 2 mm. long. The latter are simple, septate, with a
spherical terminal cell, and from 0.08 mm. to 1 mm. long'. The terminal cell
(or sometimes two cells) is surrounded by a globule of secretion. In this
doubtless resides the ethereal oil which gives the characteristic scent to
the plant. Waters (1903, p. 290) states that the odor is stronger in plants
grown in dry, sunny places than in those grown in shade, and that it is
changed and intensified in drying the leaves. By "distilling with steam"
a "considerable quantity of the partly dried ferns, . . . two or three
drops of oil were obtained ... It had a rather disagreeable odor, but
when a drop or two of a solution of the oil in a large amount of ether was
put on paper and the ether allowed to evaporate, a very pleasant reminder
of 'new-mown hay' resulted" (Waters, 1903). One bottle of fronds pre-
served in 50 per cent alcohol has retained the odor strongly, and it adheres
very persistently to hands or clothing after the alcohol has evaporated.

The leaf-shoot, several times referred to above, attracted my attention
in the fall of 1901. It is the stem which comes off from the base of the petiole
(fig. 4). About 20 per cent of the leaves bear such shoots, the remaining
80 per cent showing no trace of them whatever. The shoot arises on the
side of the petiole at a very early stage of development, but I was not able
to find its relation to the sectioning of the segments of the leaf. I believe

vSPOROPHYTE. 29

the shoots arise often after the section lines are obliterated. They lie
above (ventral to) the stdmatic line of the petiole. At maturity the shoot
is from 3 to 8 mm. (average of 10 ==5. 4 mm.) from the rhizome. It is in
all respects a stem. Its tubular stele is attached nearly at right angles to
one margin of the trough-like vascular bundle of the petiole (fig-. 82).
The stele usually appears truncate at its point of origin, but it may be slit
on its upper side, giving a trough-like shape (figs. 93-96). I have never
found any commissural strand connecting the shoot directly with the stem.
as noted by Gwynne-Vaughan (1903) in other solenostelic ferns. The
inner parenchyma of the petiole is continuous with the medulla of the shoot.
Occasionally two shoots arise on opposite sides of a single petiole; they are
then either opposite or as much as 2 mm. apart. The shoot may remain
dormant as a mere papilla for two or three years, or its growth may be
extremely slow, or it may from the first nearly equal in size the leaf from
which it springs, or the leaf may fail to develop (owing to injury by fungi,
etc.) and the shoots alone continue. Ultimately the leaf-shoots produce
normal rhizomes, as shown by one which I measured in March, 1905. It
was 12 cm. long, 1.5 mm. in diameter at its origin and 2.6 mm. at 4 cm.
from its origin. During previous seasons it had borne three leaves, and
two more were ready to develop in 1905. The apex had just forked.

The descriptions of leaf-development given by Sadebeck (1873, 1874,
1878, 1898), Kny (1875), and Campbell (1887, 1895, 1905) contain some
points which I can not interpret, and the leaf of Dcnnsttcdtia differs in its
development from that of the Hymenophyllacese as given clearly by Prantl
(1875, a, />). The leaf takes its origin from a deep four-sided prismatic
cell in the apex of the stem (figs. 103, 109, 118, 121, 128). The cell is
sometimes recognizable by its large size in the fourth segment from the
stem initial (fig. 109). It seems reasonably certain that not every stem
segment gives rise to a leaf, nor even two out of every three segments.
The location of the leaf -initial in the segment also varies much, according
to the size and rapidity of growth of the stem. It is usually near one mar-
gin of the segment (figs. 103, 109), and may lie very near the stem-initial
until many divisions have occurred in it (figs. 128, 129). The leaf-initial
in its earliest stage extends into the stem as far as the future boundary of
medulla and plerome. Its divisions are different from those of the neigh-
boring cells — a fact which is related to the formation of the leaf -gap. No
definite order could be discovered in the early divisions of the leaf-initial.
It maintains its four-sided shape for some time (figs. 118, 121), then is cut
obliquely (figs. 121, 122-127) and becomes tetrahedral. After forming a
few segments on three sides, the initial ceases to form segments on one of the
three sides and it becomes ' 'two-sided. ' ' Meanwhile, all of the cells derived
from the primitive leaf -initial are elongating and forming" a papilla on the
stem apex, at the tip of which the definite apical cell of the leaf is situated
(fig. 70). This cell is wedge-shaped (figs. 123, 130) and very broad across

30 STRUCTURE AND LIFE-HISTORY OF HAY-SCENTED FERN.

its exposed face. As segments are cut off from it alternately on the two
sides, it itself becomes narrower, and finally very slender (figs. 131, 135).
Each segment divides first on the ventral side by a radial anticline which
cuts off a narrow cell running from the periphery to the center of the
leaf-rudiment (fit^s. 132-134, 140). A similar wall near the dorsal margin
of the segment meets the first nearly at right angles, close to the center
of the leaf -rudiment. The two narrow cells may be called sections (Johnson,
1898) and the walls secfion-a'a/ls. The remaining triangular portion of
the segment is a primary marginal cell. The segment enlarges and a wall
parallel to the segment walls (transverse anticline) divides the primary
marginal cell into two equal secondary marginal cells (figs. 132-134). In
each of these two new section-walls occur, one ventral, one dorsal. Thus
there is a regular alternation of section-walls with a halving of the mar-
ginal cells. At least six or seven section-walls are formed in the region
which is to become petiole or rachis. The ultimate marginal cell is then
cut across by a periclinal wall (figs. 142, 143). Its specific activity is ended
and its outer portion breaks up into epidermal cells. The relation of the
sections to the inner tissues of the petiole and rachis (vascular tissues, etc.)
differs in different places, and was not worked out in detail.

No sooner has the leaf- rudiment become a conical projection on the stem
than it begins to grow more actively on the dorsal side — to bend over
toward the stem-initial — to become circinate. The initial cell of the leaf
lies with one point (in cross section) towards the stem initial and the ven-
tral (upper) surface of the leaf, the other point in the dorsal (lower) surface
of the leaf.

The circinate vernation comes about through the rapid growth in thick-
ness of the dorsal sections of the segments. The cells divide about twice
as rapidly as in the ventral sections, and are larger (figs. 136, 139). A
little later, divisions occur on the ventral side to about the number of those
on the dorsal, but the curved position is maintained by greater elongation
of the dorsal cells. Finally, when the leaf unfolds, the ventral cells elongate
to equal those of the dorsal surface.

The activity of the initial cell of the leaf is, as in most ferns, limited.
After the rudiments of five to eight pairs of pinnae are visible and about
three pairs of segments are cut off in advance of any visible differentiation
into pinna\ the initial ceases to exist as such. Probably it simply begins
to divide into sections as a segment would do. Fig. 138 shows a leaf-tip
at this stage, where the initial has divided twice in succession on the same
side. There is no evidence at all for a "transverse" division of the
initial. After the initial is lost the leaf-apex is occupied by a group of
marginal cells (figs. 137, 144) which grow and section and divide into
halves for a long time. Since it is probable that each segment of the
initial (while it lasts) develops a single pinna in the region of the lamina,
we may say that the lowest eight to eleven pairs spring from as many

SPOROPHYTE. 31

segments. The remaining" 22 to 40 pairs of pinme come out after the
single initial is lost. In either ease their actual history is the same.

The apical growth of the leaf, with or without a single initial, gives
rise direct!)' to a slightly flattened and circinately curved rod of embryonic
cells (figs. 140, 141, 144). Each margin is occupied by a row of marginal
cells (fig. 137). Where a pinna is to develop, about six consecutive
marginal cells elongate to form a papilla. By sectioning and halving they
rapidly increase in number of cells and mass of tissue. The apex of the
papilla and its manner of growth are exactly like those of the tip of the
leaf after the loss of the initial cell (figs. 145-147). On the sides of this
protuberance similar outgrowths form the pinnules, and on the pinnules,
in a similar manner, the lobes of the pinnule arise, and on these again, in
like manner, the ultimate crcnations of the leaf-margin (figs. 5, 148, 149).
From the inner ends of the oldest sections in each part of the leaf the
vascular tissues are derived (fig. 149). In every case also there are in-
active marginal cells between those groups which grow out to form pinnae,
pinnules, lobes, and teeth. These sluggish cells ultimately give rise to
the tissues along the raehis between the pinme, or along the ribs between
the pinnules, or in the notches of the pinnules (fig. 148). In the lamina
proper, away from any vein, the sections of the marginal cell are broad
and shallow, extending only half the thickness of the leaf (fig. 150). Each
of the sections is halved parallel to the surface of the leaf. The outer half
is an epidermal cell; the inner half remains single or divides again in the
same plane and forms parenchyma (fig. 150). The ultimate marginal
cells constitute the margin of the mature leaf. Stomata are formed while
the epidermal cells are still polygonal in outline, and while the leaf is un-
folding. A curved wall cuts out the stoma mother-cell on one side of a
young epidermal cell (figs. 152, 153). The mother-cell becomes oval and
is divided longitudinally into the two guard-cells.

With 20 to 25 pairs of rudimentary pinnae, no stomata, no air-spaces in
the parenchyma, and no signs of fructification, the leaves emerge from
the soil. It would be, therefore, quite a mistake to suppose that in Dcnn-
st&dtia piindilobula the unfolding of the leaves "consists merely in an ex-
pansion of the leaf with comparatively little cell division' ' (Campbell, 1895,
p. 325; 1905, p. 333), in spite of the rapidity with which the unfolding
takes place. One-third to one-half of the blade of the leaf must be made
outright during this time. In eastern Pennsylvania and Maryland the leaves
appear above ground in the latter half of April (Cockeysville, Maryland,
April 21; Oxford Valley, Pennsylvania., April 29, 1905). By June 4 (Loch
Raven, Maryland, 1905) spores are nearly mature. At first the petioles,
green in all the aerial part and clothed with white hairs, elongate and unroll.
Then the leaf spreads out from below upward. A comparison of figs. 2 and 5
of the mature leaf and fig. 152 of a pinnule 3 mm. long from an unrolling
leaf will give an idea of the change that goes on. In fig. 152 the stomata

32 STRUCTURE AND LIFE-HISTORY OF HAY-SCENTED FERN.

are just developing". The pinna from which this was taken was 1.5 cm.
long, and unrolling.

While it is unfolding (May 4, 1906, Baltimore, Maryland), the fertile leaf
acquires its sori. In origin the sori are strictly marginal. At the point
where a sorus is to develop, a marginal cell of the lamina, at the tip of a
rudimentary veinlet, after giving rise to a mass of vein and lamina cells,
grows out into a short, rounded papilla (figs. 151, 158; cf. also fig. 63o in
Sadcbeck, 1898). This papilla is the rudiment of the first and central spo-
rangium. Neighboring cells elongate to form a mound, the placenta. New
sporangia at once begin to develop around the first one. Meanwhile, about
four or five cells removed from the central marginal cell, the leaf-tissues
begin to rise up in a ring (indusium) around the placenta (fig. 158). The
ventral (upper) part of this ring soon becomes much thicker than the
opposite side — as thick, indeed, as the lamina itself. Vascular tissues, also,
are formed for a short distance into this lobe (fig. 155).

As growth proceeds, the sorus reaches its ultimate position on the under
side of the leaf. One is found on the lower outer venule of each lobe of
each pinnule (fig. 5). At maturity the indusium is circular and cup-
shaped. One side of it is continuous with the margin of the leaf (fig. 155);
elsewhere it rises abruptly from the surface. In its lower parts it con-
sists of inner and outer epidermis, with a few parenchyma cells and air-
spaces between. Here stomata occur both within (fig. 120) and without
(fig. 119). The epidermal cells are wavy-margined (fig. 115). The
indusium tapers above to two cells, then to one cell in thickness. On its
sides and margin it bears hairs, both glandular and acicular. The margin
is irregular. In the bottom of the indusium cup and on the side nearest
to the base of the pinnule is a low, rounded placenta. It is covered with
epidermis, beneath which is a layer of parenchyma, and then a group of
short scalarifonn tracheids (fig. 155). These last constitute morphologic-
ally the end of the neighboring venule, which appears to terminate under
the sorus just />nw/</thc placenta. The apparent ending is really a branch
in the indusium. From the placenta arise a few paraphyses (figs. 169,
173) and a number of sporangia. The paraphyses are obtuse hairs, com-
posed of about three cylindrical cells each.

The sporangia arise in centrifugal succession from superficial cells of
the placenta. The mother-cell bulges out considerably and is cut by an
oblique wall from the middle of its base to one side of its summit (fig.
159). A second oblique wall strikes across from one side of the summit
to the first wall (fig. 160). A third wall, striking both of the preceding,
leaves an upper cell with a spherical outer surface and a triangular pyra-
midal base. Three more divisions follow in the upper cell, parallel to the
first three dig. 157). Then a transverse wall cuts across the top, leaving
a central tetrahedral cell (the primitive archesporium) surrounded by four
wall-cells digs. 154, 156). The three lowest and earliest-formed cells

SPOROPHYTE. 33

divide only transversely, and give rise to the three rows of cells of the
stalk (fiy. 176). The four wall-cells form the walls of the sporangium.

The archesporial cell at once cuts off four tapetal cells (figs. 161, 162),
one on each of its faces, beginning" on the sides. Each tapetal cell divides
into four by anticlinal walls. The archesporial cell then divides nearly
vertically into two, and the tapetum divides periclinally into two layers
(fig-. 163). The exact relation in time of the divisions of archesporium
and tapetum is not constant (figs. 164, 166). From this time onward the
archesporial cells divide exactly synchronously, though in different planes,
first into four (figs. 165, 166), then into eight, and finally into sixteen.
As soon as the sixteen spore mother-cells are formed the tapetum begins to
degenerate (figs. 167, 170). Its walls dissolve, and the cytoplasm forms
a vacuolated mass, but the nuclei persist until the spores are acquiring
their definite walls (fig\ 168). The spore mother-cells, at first in a solid
mass (fig. 167), enlarge rapidly and separate from one another. Their
nuclei especially increase to a relatively great size (fig. 170). The chro-
matin now lies in innumerable fine granules. Then a long, fine, and
tangled chromatin thread is organized, and synapsis ensues. This must
be a lengthy stage, as it is frequently and easily found. Emerging from
synapsis, a heterotypic division occurs. The two resulting nuclei at once
divide again, and spindle fibers are formed between all four of the daughter
nuclei. Across the fibers cell-walls are laid down. The spores remain
together in fours until their walls are thickened. They separate before
the final sculpturings are formed on the outside. Thus the spores always
show a tetrahedral angle. The sculpturings are irregular lines and lumps
of brown substance (figs. 174, 175). All stages in the formation of
sporangia and spores may be found in a single collection of material from
newly unfolded fronds (e. g., Loch Raven, Maryland, at base of a steep hill
facing- north, June 4, 1905). Meanwhile the wall of the sporangium has
also reached its mature size and structure. The sides consist of very thin,
broad cells, about eleven on each side. The right and left sides are nearly
identical (figs. 171, 172). The annulus runs from the stalk up one edge
and over the top of the sporangium and about one-quarter of the wTay
down the other edge. Its cells, 18 or 19 in number, are cubical and
are heavily lignified on all except the outer walls. The mouth of the
sporangium, or point of rupture, lies between two long, narrow, slightly
thickened cells (sometimes two such cells on one or the other lip) about
midway between the end of the annulus and the stalk. Similar long', narrow
cells, three to five in number, but with thin walls, fill up the space above
and below the mouth. At maturity the whole sporangium and contents
dry out, and the spores are hurled away as described for another fern by
Atkinson (1894). The annulus tends to straighten, and it does so, slowly
opening the mouth of the sporangium and tearing the walls nearly straig'ht
across their whole width. It bends far .around backward until the head

34 STRUCTURE AND LIFE-HISTORY OF HAY-SCENTED FERN.

of the sporangium nearly touches the stalk. Meanwhile some spores have
fallen out, others remain in each half of the sporangium. Then suddenly
the annulus springs back to its former place, throwing its load of spores
several centimeters. Straightway it begins slowly to bend back again,
and repeats the operation. The springing of the thin walls during both
the extension and the recoil of the annulus throws out several spores.

GAMETOPHYTE.

Spores sown on moist micaceous earth about Octobers, 1905, had mostly
germinated and formed protonemata by the 25th.* About October 1,
1905, also, fresh male and female prothallia and young plants in various
stages of development were found in a ravine near Baltimore, Maryland.
In germination the spore-coat bursts at the tetrahedral angle. The intine
bulges out, and a cell with many chloroplasts appears (fig. 177). It imme-
diately sends out a rhizoid which is separated by a wall from the cell.
Some of my cultures became temporarily dry at this stage, and the first
rhizoid died. Another was soon sent out from a point farther up on the
same cell (figs. 191, 193). The protonemata assumed a great variety of shapes
(figs. 178-199). There were from one to six cells, in linear series. The
basal cell was sometimes very long and slender (figs. 184, 196), or rarely
broader than long (fig. 182). The basal cell was usually the longest of all,
but not always (fig. 197). The growth of the protonema is by transverse
divisions in the apical cell. Rarely an intercalary division occurs in a
longitudinal plane.

Sooner or later the terminal cell divides longitudinally in half (fig. 189).
( )ne half enlarges more than the other, causing the partition to become
oblique. Then an oblique wall cuts the larger cell, striking the next
earlier wall nearly at right angles (figs. 185, 196). The result is a two-
sided apical cell, which continues to divide parallel to its two sides for a long
time (figs. 187, 188, 190). The young gametophyte now becomes broader
at the apex (figs. 194, 195). Soon the cells on either side of the initial
outgrow those just behind it, and the prothallus assumes its cordate shape
(figs. 199, 200). The initial is then changed by a transverse wall across
its posterior end (fig. 203). After this it divides into two longitudinally,
and we can henceforth recognize only a group of marginal initials (fig1.
200). Each of these cells divides on three sides, viz, two lateral and one
posterior (figs. 202, 204-207, 214).

*My cultures were sown in 3-inch to 6-inch flower-pots and pans on micaceous soil
dug from deep down in a newly exposed bank of earth. Ripe leaves were laid on the
pots, covered loosely witli papers, and allowed to dry and shed their spores. Or, the
de"bris of dried fertile leaves was sown. The pots and pans were never sprinkled, but
were kept standing in i to 2 cm. of water, covered with glass plates, before a west
window of the Johns Hopkins Biological Laboratory.

GAMETOPHYTE. 35

Meanwhile many rhizoids are reaching' from the under surface of the
prothallus into the earth (fig. 231). Hitherto all of the divisions men-
tioned have been vertical, and there is but one layer of cells. When, as
in female plants, a "cushion" is formed in the middle of the prothallus,
divisions parallel to the surface occur in the posterior segments of the
initials (figs. 202, 205-207, 214). The upper half of the segment may-
divide once or twice in this plane, giving- rise to two or three layers of
cells. The lower half is responsible for most of the bulk of the cushion
and for all of the organs which appear thereon (figs. 206, 207). In
crowded cultures wart-like outgrowths occur on the upper surface of the
prothalhis (fig. 201). They can only be considered as abnormalities.

The prothallia reach sexual maturity at five weeks and later from sowing
of spores. They are practically always dioecious. Only three hermaph-
rodites have I seen. On one of these male and female organs seemed to
be mature at the same time. The males mature first, and they may con-
tinue to grow and bear great numbers of antheridia for five months or
more. Antheridia may appear on very small plants. I found one in a glen
near Baltimore, Maryland, in which there were four prothallial cells and
three antheridia (fig. 198). All sizes and shapes occur from this up to
those which are 5 mm. across, with many lobes, and the margins crisped like
a "curly" lettuce-leaf. Whether those which bear sexual organs at a very
early age ever develop to large size I do not know. Probably they do not.

The male prothallus is always but one cell thick; it has no cushion.
The antheridia arise at any point on the shaded side — central or marginal
(figs. 227, 228). In large "curly' specimens the relation to light is
clearly shown. A section may show an S-shaped portion of prothallus.
Supposing the light to come from the tipper edge of the page, the anthe-
ridia will be found on the lower side of each transverse bar of the S.
Two of these will be the morphologically lower surface and one the upper.
At transition points two antheridia may be seen on opposite sides of the
same cell !

The antheridium arises from the prothallial cell as a papilla, which is
soon cut off by a basal wall (fig. 225). It differs in appearance from a
young rhizoid in having many chloroplasts, though the rhizoid rudiment
may have two or three. In preserved material this difference is not evi-
dent. The papilla enlarges and is cut in two by another wall parallel to
the first (fig. 226). The first cell is the stalk-cell, and undergoes no
further division; the other is the antheridial mother-cell. The next two
walls are those commonly described for fern antheridia, viz, first a dome-
shaped wall parallel to the outer wall of the mother-cell (figs. 228, 232);
second, a circular wall at the summit of the outer cell to form the lid.
The body of the antheridium now consists of a cylindrical wall-cell, a cir-
cular cap-cell, and a large, dense central cell. The central cell is devoid
of chlorophyll. It divides at first vertically (fig. 227), then in three

36 vSTRUCTURE AND LIFE-HISTORY OF HAY-SCENTED FERN.

planes at right angles to one another, and at last irregularly (fig. 230), to
produce 32 sperm mother-cells. These separate and round off, and in
each one a spiral spermatozoid develops. As the central mass enlarges it
pushes itself down into the center of the stalk-cell until it reaches the
original basal wall (fig's. 230, 237). Thus the stalk-cell becomes ring-
shaped, though at first it was a disk (figs. 226, 229). At maturity the
contents of the antheridium swell up by absorption of water, the cap-cell
is ruptured irregularly, and the sperm mother-cells escape. After 45 to
50 seconds the mother-cell wall bursts and the spermatozoid swims rapidly
away with a steady rotary motion, bearing a granular vesicle at the pos-
terior end. From cultures of male prothallia kept free from surface-water
it is easy to get numbers of spermatozoids by mounting the prothallia in
water. Mr. W. D. Hoyt found them to be distinctly attracted by a malic-
acid solution of suitable strength. The body of the sperm makes about
two and a half coils. The anterior end is more slender and more closely
coiled.

TABLE 7. — Development of antheridium.

<Pro.thallial cell
.Stalk cell
Antheridial cell( /dome-like l^walfcell

Body cell/ outer cell )\ cover cell

inner cell — 32 sperm mother cells

Female prothallia are always cordate in shape and bear their archegonia
on the under side of a central thickening or cushion. They begin to form
sexual organs when still narrower than long, five or six weeks from s<>\v
ing of spores. If not fecundated they continue to grow larger and broader,
and produce many archegonia in succession over all the central lower sur-
face. The upper surface is at first flat, but in old females the margins
may become reflexed dorsally, and the plant forms a broad, erect funnel,
open on one side. The largest in my cultures are 5 to 7 mm. tall and 6
to 9 mm. across. I found one out of doors 7 mm. wide and 4 mm. long.

The archegonium develops from a cubical superficial cell of the pro-
thallus, near to the initial cells. This cell may be a semi-segment, involv-
ing half the thickness of the prothallus (fig. 202), or it may be simply
the outer part of a semi-segment (figs. 206, 214), according to its point of
origin on the prothallus and the size and thickness of the latter. In any
case the definitive archegonial mother-cell first cuts off a thin superficial
cell, the neck rudiment (fig. 202). Then on the opposite side a similar
basal cell is separated, leaving a large central cell (figs. 206, 207). The
basal cell divides crosswise into four and forms the innermost part of the
wall of the archegonium (fig. 208). The neck rudiment similarly divides
crosswise into four (fig. 207). Its first cleavage wall is parallel to the
longitudinal axis of the prothallus. Now the central cell enlarges and
pushes out the four neck-rudiments. As the latter project more and more

GAMETOPHYTE. 37

from the surface of the cushion they divide each into a row of cells (figs.
208-213). The neck, therefore, consists of four rows of cells (figs. 213,
220-222), two anterior and two posterior. The divisions always occur in
the uppermost or next to uppermost members (fig. 210). At maturity
the neck bends over strongly away from the growing point of the pro-
thallus (figs. 207, 211, 219). In relation to this we find in each of the two
rows of cells of the neck on the convex side two cells more than on the
concave side (4 and 6, or 5 and 7).

Meanwhile the central cell has cut off a "neck canal-cell" (figs. 209,
210), which pushes up in the axis of the neck. It acquires two nuclei
(fig. 215), rarely three (fig 218). Another division in the central cell
cuts off the "ventral canal-cell," lying at the base of the neck (figs. 216,
217, 218). The large remainder is the egg-cell. As the archegonium
matures the neck enlarges and becomes swollen near the end (fig. 219).
The canal-cells degenerate into an amorphous mass, the central parts of
which stain deeply with haematoxylin. At this time also a distinct venter
wall is formed around the egg by divisions in all the prothallial cells sur-
rounding it (figs. 219, 223, 224). To recapitulate, the archegonium as a
whole is made up from two sources — the neck, canal-cells, egg, and four
basal cells of the venter are all derived from the original cubical arche-
gonium mother-cell; the side walls of the venter are derived from all the
neighboring prothallial cells.

TABLE 8. — Development of archegonium.

.neck cell 4 neck cells— 4 rows of neck cells

Mother /
cell \ /neck canal cell;-neck canal cell, 2 or 3 nuclei

^interior cell-r-central cell/

\ .xventral canal cell

central cell-^

ovum

basal cell— 4 basal cells — base of venter wall
Surrounding cells of prothallus form side walls of venter.

When the archegonium is wholly mature, the uppermost four or eight
neck-cells break apart, leaving a wide-open mouth (figs. 220, 224).
Through this a transparent mucilaginous substance exudes, and may
stream out for a distance several times the length of the archegonium
(fig. 224). In this substance spermatozoids gather in great numbers.
As a sperm enters the mucilage its movements become slower, and it
changes from a short, stout helix to a long, slender one with more turns.
The vesicle attached to its posterior end is twisted off by the resistance of
the mucilage and floats away. The sperms swarm into the neck and
make their way down towards the egg, which becomes pointed as though
reaching out to receive them (fig. 223). The lower part of the neck is so
constricted (fig. 222) that the sperms have to become nearly straight to
get through, but many succeed in doing so.

.STRUCTURE AND LIFE-HISTORY OF HAY-SCENTED FERN.

The remains of unsuccessful sperms are left after fecundation as a deeply-
staining' cap over the top of the egg-cell. The fertilized egg- rounds off
and comes to rest, with a large central nucleus and nucleolus (fig. 224).

About half of the mature prothalli, with apparently perfect archegonia,
will actually be fertilized when mounted on a slide with mature males. In
such eases not only is the receptive archegonium filled with sperms, but
many older archegonia, up to a dozen on a single plant, including such as
are brown with ag'e, are quite as eagerly crowded into by numbers of
sperms. Only once, however, have I known two embryos to appear on
one prothallus. Young archegonia continue to develop on the fertile
prothallus for a time. But after the embryo is fully established (i.e.,
octant stage), sexual organs cease to develop. Fertilized eggs were found
about 7 days after my cultures had been flooded with water. In 16 days
many embryos were visible with a hand-lens.

THE YOUNG SPOROPHYTE.

The first cleavage-plane (basal wall) in the fertilized egg includes the
axis of the archegonium, and lies transversely to the axis of growth of the
prothallus. It divides the egg into anterior and posterior halves (see
below). The second (quadrant) wall passes horizontally and at right
angles to the axis of the archegonium. In each quadrant, then, a vertical
wall is formed at right angles to the two preceding, dividing the embryo
into octants. These octant walls do not correspond in the different quad-
rants, but the octants are from the first unequal in size (figs. 233, 234).
vSupposing the prothallus to lie before the observer with cushion down-
ward and the notch on the farther side, we may speak of right and left,
anterior and posterior, upper and lower portions. The fate of the octants
may be stated thus:

1. Anterior upper right octant=Stem initial )

, or vice versa:.

2. Anterior upper left octant ^Irregular

3. Anterior lower right octant — \

4. Anterior lower left octant = i First

5. Posterior upper right octant =: j

6. Posterior upper left octant = i

7. Posterior lower right octant=Irregular | Qr ^ vers(e

8. Posterior lower left octant —Root )

It must not be supposed, however, that this arrangement is invariable.
On a prothallus with two embryos one has the root-intitial in octant 8, the
other in 7. Octants 2, 3, 5, 7 are smaller than 1, 4, 6, 8. The first
division in 2, 4, 6, 8 is parallel to the basal wall and near to it. In 8
the succeeding- divisions are parallel to the other primary walls, and then
to the curved outer wall. The resulting tetrahedral central cell is the
root initial. It continues to divide in the way which is characteristic for
roots (7. v.) (figs. '235, 236).

THE YOUNG SPOROPHYTE. 39

The two posterior upper octants (5, 6) divide irregularly into a mass of
polyhedral cells, the foot. Those in contact with the base of the arche-
gonium become closely applied to the wall, and the boundary between
prothallial and embryonic tissues is often difficult to determine. Neigh-
boring" cells of the two anterior upper octants are also involved in the for-
mation of the fully matured foot (figs. 235, 236, 246).

Octants 2 and 7 divide irregularly and serve only to fill up their respec-
tive corners of the embryo plant. Octant 1, after cleavages mostly parallel
to the basal and qtiadrant walls, ultimately gives rise to the stem-initial,
lying close to the octant wall, /. c., near the median line.

The lower anterior octants (3 and 4) elongate together in a horizontal
direction (fig. 255). They unite at their anterior ends to form a group of
marginal initials for the first leaf (cotyledon). This leaf , therefore, never
possesses a single apical cell.

As the leaf grows out, the whole anterior (epibasal) half of the embryo
elongates, carrying forward both stem-initial and leaf. The plantlet lies
horizontally. It consists of a short cylinder with root-initial at one end,
leaf -initials at the other end, a dorsal papilla near the middle, in which is
the stem-initial, and back of this a large dorsal protuberance, the foot,
buried in the prothallus (fig. 246) . The whole is surrounded by the greatly
enlarged archegonial wall, the calyptra. The latter has become two or
three cells thick all round (figs. 246, 255). Soon the leaf bursts through
the calyptra and bends upward through the notch of the prothallus, and
the primary root extends downward. The new plant is now independent,
but the prothallus does not disappear until two or three leaves are formed
(fig-. 267).

The young stem grows almost horizontally for 1 or 2 mm., increasing'
in diameter and complexity of structure until about five leaves have been
formed. The stem then forks. The plant now differs only in size and
sterility from the adult. The primary root grows about 5 cm. long, is
slender, and has the structure of an adult rootlet (</. ?'.). It is not
branched, but has copious root-hairs. Adventitious roots arise in rapid
succession, being sometimes as many, sometimes twice as many, as the
leaves (fig'. 256). They are like adult roots, only smaller. I have one
specimen whose root-cells are densely filled with fungous hyphae, after the
manner of the adult mycorhiza. The prothallus is also infected, but there
is no connection (internally at least) between the fungous masses in sporo-
phyte and gametophyte.

The sporeling' stem is short and cylindrical (figure 269). It is clothed
with an irregular epidermis. The cortex, parenchymatous below, merg-es
gradually into that of the adult region. The stele of the primary root is
continuous with that of the stem. There is no line of demarcation on the
ventral side, but dorsally a prominent angle of vascular tissue projects
toward the foot (figs. 246, 247). Between this point and the insertion of

40 STRUCTURE AND LIFE-HISTORY OF HAY-SCENTED FERN.

the first leaf the stem is protostelic. It contains a solid core of scalariform
tracheids (xylem), surrounded in succession by conjunctive parenchyma,
phloem, pericycle, and endodermis (fig. 238). At the exit of the first
leaf-trace there is a depression containing parenchyma on that side of the
xylem (fig. 239). Farther up the xylem extends around so as to inclose
this parenchyma. In the parenchyma, even before the exit of the second
leaf -trace, phloem and conjunctive parenchyma may be recognized (fig.
240). The second leaf makes a similar gap in the xylem. Between the
second and third leaf-gaps there appears in the midst of the central phloem
a group of large cells identical in appearance with the outer pericycle (fig.
241, if>). At the exit of the third leaf-trace the outer pericycle becomes
continuous with the central tissue just mentioned, through a gap in the
xylem and phloem (fig. 256). The gap closes again without any dipping:
in of the endodermis. Below the fourth leaf there appears a thickened
band (Caspary's band) on the cross wall between two parenchyma cells at
the center of the stele (fig. 242). In the next section (10 p- higher) there
is a line of five cells whose intermediate walls bear the thickened band
characteristic of endodermis (fig. 243). Two sections higher (20 /*) there
is a ring of endodermis at the center of the stele, inclosing one scleren-
chyma cell (fig. 244). The ring enlarges rapidly and parenchyma cells
appear beside the sclerenchyma (fig. 245). The solenostelic structure is
established. The fourth leaf-gap is like the third, with only a very slight
dipping inward of the outer endodermis (fig. 257). Only at the fifth or
sixth gap, /. e., above the fork of the stem, is there a continuity established
between inner and outer endodermis and between medulla and cortex, as
in the adult plant. The above description is of an average case. The
exact position of each stage differs according to the size of the individual
sporeling. The whole course of development is remarkably identical with
that described by Boodle (1901) for the early stages of Aiiciinia fihyllitidis.
Dennstcedtia stops at the solenostelic stage, while Aneimia goes on to
become dictyostelic.

To the practiced eye the first leaf of a young fern is often sufficient to
distinguish the species. In any case the third, fourth, or fifth leaf will
bear undoubted specific characteristics. The first leaf of Dennsteedtia
piiuctilobnla is usually two-parted, with each part bifid at the apex (figs.
259, 267). In more slender examples the two lobes are narrow, elliptic,
and entire. I have seen two cases where the leaf bore but one elliptic
entire bit of lamina. The average leaf measures 0.32 cm. to 0.38 cm.
across. Its venation is simply forked. In the bud it is folded over at the
tip in involute manner, but could hardly be called circinate. The same is
true of the rudiment of the second leaf. The second leaf is also broad
and lobcd. It has three or four main lobes, each bifid or emarginate at
apex. It is larger than the first, being 0.46 to 0.81 cm. in width and 0.4
cm. or less in length (figures 260, 261). The third leaf is pinnately

THE YOUNG SPOROPHYTE. 41

divided, with a comparatively broad, winged rachis. There are one or
two pairs of pinnules and a terminal portion, all of which are lobed and
erenated (figs. 262, 263). The fourth and succeeding leaves are pinnate
like the third, but with more pinnae (figs. 264, 265). All of these early
leaves are broadest at the base and they vary from deltoid to broadly
lanceolate in shape. But in outline of the pinnae and pinnules the third
and fourth leaves are exactly like the mature leaf. They are thin and
fragile, consisting only of upper and lower epidermis (figs. 254, 266) and
one layer of spongy parenchyma (fig. 253). Stomata are numerous in
the lower epidermis of each leaf, especially on the pinnate leaves. A few
stomata occur scattered over the backs of the petioles (fig'. 268). The
margin of the leaf is strengthened by long, narrow, indurated cells imder-
lying the epidermis (fig. 251).

The first two leaves are devoid of hairs of any kind. Hairs begin to
occur on the third leaf, but the fourth shows three kinds of trichome struc-
ture— glands, moniliform hairs, and acicular hairs. A few glands (fig.
258, w) occur, thinly scattered on the upper and lower surfaces of lamina
and rachis, but they are more plentiful on the petiole. Each one consists
of one to three large, swollen cells. They probably represent the glandular
hairs of the adult. Moniliform hairs (fig. 258, w) consist of three or four
cells, each of which is broader above and narrower below. They lie
appressed to the leaf-surface. They are plentiful on both surfaces of the
lamina and rachis, but there are none at all on the petiole. Acicular hairs
like those of the adult leaf are plentiful all over the fifth leaf, and on the
stem apex. They are four to seven cells long (commonly 4, 5, or 6),
thick- walled, and curve outward from the surface of both the lamina and
the petiole.

The mature petiole of the early leaves is slender, flattened above and
rounded below (fig. 248). Under an uneven epidermis there is a cortex
composed of two or three layers of large, thin- walled cells. In this are
large intercellular spaces. A well-defined endodermis demarcates a cyl-
indrical vascular bundle. In this is a stout transverse band of xylem,
surrounded by phloem and a single layer of pericycle. The xylem con-
sists of narrow spiral and scalariform tracheids.

As stated above, the first leaf is derived from two octants of the embryo,
not from one alone, and grows always by a group of marginal initials.
These undergo sectioning and halving" as in the adult the leaves (fig. 250).
The second leaf arises from the stem-tip. Its development has not been
followed.

I have not determined how long it takes to obtain mature plants from
spores. Forked stems are found after about one year. In some cases
certainly another season intervenes before maturity is reached. Probably
they never fructify before the third or fourth summer.

42 STRUCTURE AND LIFE-HISTORY OF HAY-SCENTED FERN.

DISCUSSION OF RESULTS.

We have restricted ourselves thus far to mere description of the develop-
ment of Dennstcsdtia punctilobula. It remains to point out in order some
general considerations suggested by the investigation.

TAXONOMY.

Whether our fern belongs in the Cyatheaceae or the Polypodiaceae should
not be difficult to decide. The principal differences between the orders
may be shown by a table:

CYATHEACEAE. POLYPODIACE^:.

Annulus a complete circle, oblique. Annulussurrounding three-fourths of spo-

Antheridium cover multicellular. rangium, symmetrical.

Cushion of prothallus with bristles. Antheridium cover unicellular.

Broad cells in continuous series in the rhi- Cushion without bristles.

zogenous line. Rhizogenous cells separated by smaller

cells.

In all of these points the plant now under discussion agrees with the
Polypodiaceae. But Bauke (1876) states for Dicksonia nibiginosa that
its antheridium is cyatheaceous. Moore (1857) and recent writers place
D. rnbiginosa Kaulf. and Dicksonia punctilobnla Willd. in the genus Dcnn-
st&dtia. Gwynne-Vaughan (1903) further shows that D. rubiginosa has a
complicated solenostelic stem. If Moore and his followers are correct,
the character of the antheridium must cease to stand as a distinction
between the two orders. This point is much in need of investigation
in other Dicksonias and Dennstsedtias, as well as in DavaUia, Lindsava,
Aficrolcpia, and the allied genera. Cyatheaceous root-structure I have
observed in Cibotium regale. But we need to know the arrangement of
rhizogenous cells in the other genera just named. A knowledge of these
points is especially needed for the MelanesianDennstedttaflacada, the type
of the genus. For it may yet develop that our North American fern is
not referable to the same genus with D. Jlaccida. In that case we should
have to adopt the generic name Sitobolium Desv. The point can only be
settled after a careful examination of D.flaccida throughout its structure
and life-history. The removal of our fern from the genus Dicksonia
L'Herit. is generally agreed upon, and is quite sure to stand. The use of
the name Dicksonia certainly leads to confusion, as when a recent European
writer speaks of our plant as a "tree-fern."* But further studies along
the lines indicated are needed to fully establish the position. Indeed, it is
not impossible that such a comparison would break down the feeble barrier
between Cyatheaceae and Polypodiaceas by showing a series of connecting
links.

*"Den nordamerikanischen, 2-5' hohen Baumfarn." Brick, 1897.

DISCUSSION OF RESULTS. 43

STELAR MORPHOLOGY.

Some views already published (1906) on this point may be repeated
here. The development of the seedling- stem supports the idea of Jeffrey
and Boodle of the phytogeny of the fern-stem. We first have the pro-
tostele, then the ectoploic siphonostele, and finally the solenostele. But
there is no evidence of any influence of the tissues outside the vascular
tiibe upon those inside. Each interior tissue is established before it comes
into communication with its external homologue.

HOMOLOGY OF TISSUES.

Indeed, homology of tissues can not be determined either by continuity
or by origin in the meristem. We may not homologize the medulla of
Dennstcedtia with the central xylem of Lygodium simply because both
arise from the inner ends of the segments of the stem-initial. Much less
could we identify the inner endodermis of Dcnnstirdtia with any of the
xylem of Lygodium. On the other hand, the continuous endodermal
layers of root, stem, and leaf in Dennstfedtia must be considered as one
homog-eneous tissue. But in the root-tip the endodermis arises just out-
side and the pericycle just inside the second periclinal wall in each seg-
ment. In the stem the endodermis is the result of the last (second) peri-
clinal division in a layer of plerome which also gives rise to the pericycle.
The same is true of the leaf — an organ which grows at first by a three-
sided initial, then by a two-sided, and finally by a group of marginal cells.
In Dennstfedtia pioi dilobula, therefore, tissues are homologous which have
the same structure and function, in spite of their differences of origin
(</". Goebel, 1900).

The same conclusion is indicated by the long-familiar fact that in roots
of dicotyledons the undoubtedly homologous primary tissues arise from
at least five different types of root -tip (De Bary, 1884, p. 12). Indeed,
radically different types of tip may occur in allied genera, as in Nymphcra
and Nuphar. And it is not impossible that different types may be found
in different roots of the same individual plant. The recent discussion of
stem -tips raises the same point in questioning1 the validity of Hanstein's
tissue layers (Schoute, 1903, etc.). It seems reasonably certain that Han-
stein's layers are not of very wide application. In the face of so much
evidence, also, Van Tieghem's denial of an epidermis to the roots of ferns,
monocotyls, and Nymphaeaceae becomes quite valueless. These plants
have just as real an epidermis as have others.

INDEPENDENCE OF MERISTEM AND MATURE TISSUES.

The fact is that the development and structure of the mature tissues
are to a large extent independent of the development and structure of
the meristem from which they are derived. The lowest plant with cell-
division in three planes is essentially meristematic. From such a begin-

44 STRUCTURE AND LIFE-HISTORY OF HAY-SCENTED FERN.

ning nieristcms have had a regular and orderly phylogenetic history (</.
Bower, 1889). From the meristem a more or less homogeneous paren-
chyma is directly derived. Very soon, however, farther differentiation
occurs in the older parenchyma— the parts farthest removed from apical
meristem. This differentiation has also been modified in course of time
in response to the environment and the needs of the plant. But there is
no reason to suppose that the changes in the meristem should have any
direct relation to those in the older portions. Meristem develops from a
growing" point and its structure is influenced largely by the shape, size,
number, and position of the initial cells. Mature tissues develop from
below upward, and are shaped out of meristem in response to stimuli
which come chiefly from the more mature cells. While this independence
of meristem and mature tissues seems plainly indicated by a comparison
of plants as they commonly occur, it is also capable of experimental inves-
tigation. Some of the most promising material for this purpose would be
in the genera Glcichcnia ((/. Boodle, 1901) and Lindsaya (</. Tansley and
Lulham, 1902). But this is a subject which has yet to be followed out.

Synonymy of generic names, and list of type species on strict Linnccan priority .

1. Polypodiitm — Linnaeus, 1753 — Type: P. lanceolatum (first species named.)

2. Dicksonia — L'Heritier, 1788 - Type: D. culcita (or D arborescens}.

3. Aspiciium — Swartz, 1800 — Type: A. articnlatiim (tide Underwood, 1899).

4. Dennstcedtia — Bernhardi, iSoo — Type: D.flaccida = Trichomenesfiacciaum Forst.

5. NiepArodium—MichsLUX — Date and type uncertain. Cf. Underwood, 1899, p. 265.*

6. Sitobolium — Desvaux, 1827 — Type: S, pmictilobutn.
•j. Sitolobium — J. Smith, 1841 — Type: S.punctilobum.

8. Litolobiutn — G. Kunze, 1848 — Type: L. punctilob nm.

9. Adectum — Link, 1841 — Type: A . pilosiusc ulum .

*Cf. Davenport in Rhodora, 4:is8 ff., Aug. 1902.

SYNONYMY. 45

SYNONYMY.

Nephrodhun punctilobulum Michaux, 1803,2:268.
Aspidium punctilobuluin Swartz, 1806, p. 60.

punctilobiim Willdenovv, 1810, 5:279.

Pursh, 1814, 2:664.

Poly podium pilosiusculum Muhlenberg, in Willd., 1809, p. 1076.

Willdenow, in Sclikubr, 1809, p. 125.

Dicksonia pilosiitscitla Willdenow, 1809, p. 1076; 1810, 5: 484.

Poiret, 1811, 2: 472.
Pursh, 1814, 2: 671.
Nuttall, 1818, 2: 253.
not Raddi, 1825, p. 63.
Sprengel, 1827, 4: 122.
Darlington, 1837, p. 584; also ed. u.
Hooker, 1840, 2: 264.

Bigelow, 1814, p. 254; 1824, p. 397; 1840, p. 424.
Eaton, 1836, p. 277.
Eaton and Wright, 1840, p. 224.
Torrey, 1843, 2: 502.
Wood, 1846, p. 633; 1864, p. 820.
Macoun and Burgess, 1884, p. 222.
Gray, 1889, p. 691, pi. 18.
Macoun, 1890, p. 285.
Wilson, 1897, p. 8.
Parsons, 1899, pp. 114-119.
Clute, 1901, p. 230.
Waters, 1903, pp. 64, 68, 286-290.
Eastman, 1904, p. 64.
Dicksonia pubescens Schkuhr, 1809, p. 125, pi. 131.

Presl, 1836, p. 136.

Sitobolium punctilobum Desvaux, 1827, p. 262-263.
Sitolobium punctilobum J. Smith, 1846 a; 1846 d, p. 70.

pilosiusculum J. Smith, 1841, p. 418; 1842, p. 434.
Adectum pilosiusculum Link, 1841, p. 42.
Dicksonia putictiloba Hooker, 1846, 1: 79.

Fde, 1850, p. 335.
Lowe, 1867, 8: 123-124, pi. 42.
Hooker and Baker, 1868, p. 54; 1874.
Gwynne-Vaughan, 1901, p. 85; 1903, pp. 691, 730.
Litololnum \_pnnctilobuhtni\ Kunze, 1848 (1846), p. 88.
Dicksonia punctilobula Gray, 1848, p. 628; 1856; 1858, p. 595; 1859, p. 595; 1867, p. 669, pi.

u; 1880, p. 669.

Kunze, 1848, p. 88 (written in 1846); 1850 b, p. 249.
Darlington, 1853, p. 394.
Mettenius, 1856, p. 105.
Provancher, 1862, p. 720.
Ball, 1876, p. 155.
Williamson, 1878, p. 119, pi. 45.

Eaton, 1879, 1:339-344, pl- 44-
Fowler, fide Macoun and Burgess.
Underwood, 1888; 1893.
Britton and Brown, 1896, 1: 12.
Chapman, 1860, p. 597; 1883, p. 597; 1897, p. 635.
Dennsttrdtia punctilobula Moore, 1857, pp. 97, 307.

Lawson, 1864, p. 287; iSSS, p. 233.
DeBary, 1884, p. 284.
Underwood, 1900 a, 1: 472: 1900 c, p. 122.
Britton, 1901, p. 19.
Small, 1903, p. 1 8.
Keller and^Brown, 1905, p. 14.
Dennstczdtia punctiloba Hooker and Baker.

Christ, 1897, p. 312.
Diels, 1899, p. 217.

46

STRUCTURE AND LIFE-HISTORY OF HAY-SCENTED FERN.

Authority.

3ghem Jt Douliot, 1888.

1889.

?ghem & Douliot, 1888.

?ghem A- Douliot, 1888.

3ghem & Douliot, 1888.
1889.
;ghem & Douliot, 1888.

06 06 oo" 06

oo oo oo oo

OO OO 00 OO
i-l rH F— 1 rH

*j" •*- *S ^

2-2 ° 2

3 co 3 • ^o 3 !^ 3 ^^

^o ^^ _z o
^^ flj ^3 •• U i^ "Ti ^i ^

St^^Sti fl ** 9 tf
S " ^ 'a3 w "3 § S

1

+s
.2.

a

1

-H
£

—

*

f— i

4-T

..2

-3 3

S o>

08

+^"
o

3
o

i

A- Leitgeb, 1868.
eghem & Douliot, 1888.

H j

C ^
(S c
> ff

3
cS
>

EH .H^'HPPP
c'a ~ ^ o

c3 •** c3 Q ^

>S>CP>

fl 3 S) a g S)2 ^ 3 to fl ^j

^ wJ :"^ ^ Ct :" uj ^ !*^ ^ :"^
K^ ^ 4H t^ O 2l ^ i^*" ^ P* ^

3 3

r,

3|

bt3

O "-" O O O O O

1 1 aaa

•3 - Z

a
5 1

H T3

X
§

S : : :g : : :

QJ • • • • qj
^_»..^.^_,

" ' ' 0; &H ' ' '

a> o a

"^ §

og*
"2*8

i,

O

0

0
«w
0

£

0 ... o ...

Q ... Q ...

O..+J.O

O . . *-• kJ O

:•£ O "

f ROOTS.

Median thi
gives
endoder

o

00

3

-M

. . . Several lay*

c3 • • • c3 • • •

3 d d : 3 2 °' °'

oj ' ' : » . . .

'Jl • • -32 • • •

csdo^S^o" odd code
^•0^,7., >4gt3id'd'O'd'O'd'O

• : • x

0 0

•a -3

r

£ °

S.2,

. . . IP 5* . . .

•^ 6 c — cs o _2 _£

:::*::::

. . . 0) . . . .

O OO

^

oa

W

t-u

Z

~—>

w

rd of sextant
s rise to
mis and —

O

4J

h

O
o
CM

0 ;
« 5
<u •$

•ortex

S eg

€ 5£
o go

><
• • :§£* ••••«•••

. . ."s - t- . . : . !-, . . .
... o •« 0 .... o ...

^ . -tflOu^ • • 'O>4 • •

•£ : >> 's o tj o £ :

O • c3 ^ f/3 O CG O

ortex

o :

• . • o • Y • •

. ° : <o '. '.

• X • ?„ • 0 • •
• o> *! £ X 'j • '
; +j aj g jj^ ; ;

U O^ O M

o

O
<

UJ

a > o>

:*!

o> j*

e? ^
3 ft

: <«

. o

< 4!

v. ^""-i

0 . 0

;-, O OQ" t- O O O

Q ^j t^ ^i o *^ O ^

o !

_'

o :

o2^°?too

o

Z

HH

J

1

3
O

f *•

£ c

r -

< >i

' .2

i rH

!
:

>> W fe >i

3 S3

^ a~
3 • : : :

>^ O di O —H O «- H O
^H ^ 7-1 X! -M -H <N ^H ^

•s ;

o

- ;

^I'o-So

^H

_]

UJ

U

a
3

c
C

!
0
D

>

>

Q

0 • • • •

to •

-j-
O

w
j

n

'^z
*s x

<
i

I I

5
^

JO

<i

i . . • .

0, ; • • ;
bfi . .

1 31 ii i

^ ^ *J . . .

C3>^^ !-<CS S-tCS-^

- .

~

1

~'3

':«!

s *

s J

* 0

5 *g : ; :

O-^O «iO -a:O'«3

o •

i

0

'ia O

• a? O

H

0
^

'S

e> -

2 S

•d =C

s ^

Ls

i -a

AOO^iAOOO

i; 73 73 M.2 -a -a ^3
•c :ST3 : : :

^o'oVoO • odoOVxo

.2-S.S 2'§.2.2-§-§'§.2.2'§.£

2-d

O

--

50

la

oooroooo

•d -Z T3 "3 •« -3 -3 -S

Sli

£

aj ^

' 0)

S :^.S : : :

^•O^J O w ' 'OiS'O

M '

-

o ^

• o^J

PH

^ ,

- ^

%:•.*%•.::

^ [OS ^^...o«^.o

,-, -

r3

U<5

: - ^

J *O O. ~ .S

+

Si

4 t

4
f—

H

i

0

j

JO

d
• — i

an

So

- :

-

*M

- :

— :

Name of species.

HymenophyUacete.

Hymenophyllum alatum .
demissu

Trichomanes radicans
radicans

Cyatheacl'll'.

Alsophila procera
( 'ibotium regale
Cyathea gardneri
insignis
oligocarpa
Dicksonia antaretica
sellowiana
Hemitelia setosa

Polypodiace ce,

Acrostichum colhefolium.
aureum
Adiantum capillus
hispidulum
emarginatum. .
Aspidium filix-mas
marginale
molle
molle
rivularium
serra
violascens
Asplenium celtidifolium. .

filix-ffpminn.

thelypteroides.
Blechnum occidentale. .

&

Ceratopteris thalictroides .
Do..

PC

1>

illlit

Onoclea sensi bills
Platyceriurn alcicorne
grande

CELL-LINEAGE IN FERN-ROOTS.

47

CO CO OO CO OO CO 00

co oo SB oo oo co s

OO 00 ryi?5 00 CO 00 00

FH rH CO *"* rH rH rH

o o "* o . c o 'o 'o

S oo oo ^3 "«Sj£S po — S S

3 2 O S OrH O >-l O

Q ^ ,£- 0 ^. ^0,-rQ x- P p p

^ CD -D ^ C^l CD l5*jj Q ii-*1 ^ ^ !,W ^

ig be be oo be se ^ be

s $£ .-s s ^ ss^a T s a a

o> rH o> 01 SH a> a> <D S * oi • a! •

a _•>-! J.a^hH'.CrJ.a^ S .a io .aSi -a

t^ OaT^ . "™l (*; ^ f*"j -— ' i~~i c^ "~* P^ f*^i -"^ rt i c^ ^H FH ""* ^H /"~"i

" Cy .— ' I1 ^H r«H r/i ^ 1) 1; ^ CD ^ O CD f"H

d c bj)j3 beacsibedbedbe is d>s5c> a

r> ^)^^(?i >• QQ§!?;>r?i> ^ M r^WSr^W r>

CO OO CO 00
OO OO 00 r
CO 00 CO 00

rH rH rH I— 1

o o "o 'o

Sec — co ~ ^s -5

ol o 2 o lo -
i » ( Q o o

M .Q X? Q"W

^j ^ ^x ^ *y QJ ^ •—*

bJD ofi {jfj CJ

— •*"* Q *^ '^ il- ^ -T

P H-jpSj;^ Hq S EH p M-
abea^be a bed-0
f> J^ J>pq^ ^, >^lt> ^

* Van Tieghem& Douliot infer two layers per segment in the cap wherever nothing else is designated,
t Only one cap-segment is ever formed.

«::::: '.«:::«:«: M •

x x y, *

| III
8 g 8 8

o o o o

o ::::::: :S :: :c^o : g

' ' X y, o x • X

r». • • ^ <O *, ,r, '<->•„ 1) ' <P ,/,

£ • • £r S £ o £ r; :•£ £

CJ.' ' Or^Q^ aJr/10> ^? '^ OJ

r" *-* rH L,

<1J 4) U 4)

r^ !>1 >J >,

CS cj

ojO^O "^ 58 o "cSOOC

O 4) 4) 4>
» CC GO 02

" • • o o""1 " ' >*"'""' o : o

c32°'®'->'c'l-'--l-^c<= oi.5cbO t« ;t. jj^

^ ""•"""" ^ ^ r* " >r-> ^ .J* f^

irn

x • «

:g ki S «S

. OJ r^ d) f->

:.....;.:...'. '^ -•> . o . % o :

. . . . . o>

:::•;• • • • • • c

• • • O £ u . - o

>; : : '•* • '•** £ • '• :°*B 's « 8s : ^

fS .Q^..QJQ^1' LJ Jvi ?** * O »*

i ii mi i iii i iiiil l

?S .*"..*-'>rf , — lO^ „£ Ki^J— P*i»* ^

S^^Sl^SSlSo lls5 s; g ^ -gs; fe-§

>i^ -r^o • >> >, u >i u w ^ r^^ ^_ a» Q> ^ D ,_ >»
rS :.s!o-j3cscrtooor5o j^ogo 5

rH IlM^i :^<M^(M^ r^C-lrHOq CO 3 rH 'X tf -H

a 3d

...-..' 3 O . . 3

Q O - - O • '

• . • <4H

:/• K '• Y '• : : : °

oji "3^2 *o « 7!

r» i! O rH ij *" (nO t-OOO

^o^oo ^ o o

rH H _cg rH H

:::;•: o w • • « ::

o ;c3 • « ..
cs . . re

I:::::::: 05 * : :« : :

. . . . o ' : . : : '.

:::::::::.:.. f> : "2 : :i . • . . . - .a .
::::::::::::::::: & g i ! jf : : : : : Ij § ^ : : : :

1 j :1| :1 |1|1|1 1 r3 1 3 1 |3 Jj IS !l2* 1 |

+^ QJ • *^ flj • *^ QJ "t3 QJ "^ u •. i» ti "t^ ^ "t^ • »j <1> "t" O ' ' ij ^ "t^ Q)

J-l -4^ * M . . • « -— "^ IT- « ^J *_* *4 * *-H , -r^ ^J *^ ^J • -^ ijj ^* M J_,

Oco 'OwrOcfiOcoOcC'^K O "-i O • .^ co Cc/3-Q^co O to

^o :^o:.S o^o^o^ o ^ £ ^ £ '•& o i2o 'iS^io 2 o
So :Su :S o^oSoS o ^ S x3 S :^ o So -S^o S a

jj -4-J * j_l j_,

CD CD co

QJ CD CD

_>> „ coco
cp .Q-1 ^IC*-! OO-

i-H ^

JL i ^

• • •

:=jp- :::::;:::: : a S . : : : ; : :« : : : :

i JIMM N ;l§

aQ^^^ ^S ^'c3^K* .S OOS^Wfv*

& "ggfS Si 8' : i g^Sl

S oSo -2J Sag s!lf.d

e Es*i =3 s a-o e S-S-g^

3 1 !&•§ |a -i £5 II i||

Mill ^s 1 2.1 |a

^ pS c8 2Q CO — '^Ha

^ .i-i o en £

CS CO tC

2s "5 3 d

J O -J t-

,5 s o1 £

^H P^ ^ [rl Frj

Sal- * : • : • :^°§, |3'3S || ^gp'g |-

Is IliiJillllll |l«sp!lni |||

S S3 iTj- r^ "tl^ "^ Tl^ ^ ^ "^ ^ Q. ^"* -5 ^ "^^ £ ^0 cS ^"*

C*C'tH|--:ui-J-'^^CDCQ TK fl ^ c3 O *rt ^
O £uj£ /S ? *C

P* ^H o-^"u .^o S<i3

i? s oS's ss? a-c

O*3 O*^^" j^ o:O

Oi PH 0302 O <j I-H OH

48 STRUCTURE AND LIFE-HISTORY OF HAY-SCENTED FERN.

BIBLIOGRAPHY.

ATKINSON, G. F. 1894. The Study of the Biology of Ferns.

BALL, E. H. 1876. The Indigenous Ferns of Nova Scotia, in Proc. and Trans. Nov.

Scot. Inst. of Nat. Sci., 4: 155.
BAUKE, H. 1876. Entwicklungsgeschichte des Prothalliums bei den Cyatheaceen, ver-

glichen mit derselben bei den anderen Farrnkrautern, in Jahrb. f. wiss. Botanik,

10: 49-1 1 6, pis. 6-10.
BERNHAKDI, D. I. I. 1800. Tentamen alterum rilicesin genera redigendi, in Schrader's

Journal fiir die Botanik. Gottingen, vol. 2, parts i and 2, pp. 121-136.
BIGELOW, J. 1814-1840. Florula Bostoniensis, ed. I, 1814; n, 1824; III, 1840.
BOODLE, L. A. 1901. Comparative Anatomy of Hymenophyllaceaj, Schizasaceas, and

Gleicheniaceit. II. On the anatomy of the Schizaiaceae. In Annals of Botany,

15: 359-423. June.
BOWER, F. O. 1889. The Comparative Examination of the Meristems of Ferns as a

Phylogenetic Study, in Annals of Botany, 3: 305-393

BRICK, C. 1897. Review of Wilson, 1897, in Botan. Jahrsber. 1897, 1: 470.
BRITTON, N. L. 1901. Manual of the Flora of the Northern States and Canada, p. 19.

- and BROWN, AD. 1896. An Illustrated Flora of the Northern United States, Can-

ada, and the British Possessions, 1: 12.

CAMPBELL, D. H. 1887. The Development of the Ostrich Fern, in Mem. Boston Soc.
Nat. Hist., 4: 17-52, pis. 4-7.

- 1895; 1905. The Structure and Development of the Mosses and Ferns. Macmillan,

New York: ist and 2d ed.
CHAPMAN, A. W. 1860-1897. Flora of the Southern United States; ed. I, 1860; n, 1883;

in, 1897.

CHRIST, H. 1897. Die Farnkrauter cler Erde, Jena.
CHRYSLER, M. A. 1904. The Development of the Central Cylinder of Aracea- and

Liliacea?, in Botanical Gazette, 38: 161-184, pis. xn-xv. Sept. 23, 1904.
CLUTE, W. N. 1901. Our Ferns in their Haunts. New York; 332 pp.
CONARD, H. S. 1906. Morphology of Fern Stem, in Johns Hopkins Circulars, pp. 95-

98. May.
COPELAND, E. B. 1902. The Mechanism of Stomata, in Ann. of Bot.. 16:327-364^!. 13.

June.

DARLINGTON, WM. 1853. Flora Cestrica; ed. I, 1837; m, 1853. Philadelphia.
DAVENPORT, G. E. Dicksonia pilosiuscula var. cristata, in Rhodora, 2: 220-226. Not

seen.

- 1905. Reversions and their Fluctuations, in Fern Bulletin, 13: 106-107.

DE BARY, A. 1877; 1884. Comparative Anatomy of Phanerogams and Ferns. German

ed., 1877; English trans., 1884.
DESVAUX, N. A. 1827. Prodrome de la Famille des Fougeres, in Me'm. Soc. Linn.

Paris, 6: 171-337. See pp. 262-263.
DIELS, L. 1899. Poiypocliaceai in Engler and Prantl's Natiirlichen Pflanzenfamilien,

1-4: 217-218. Sept. 12, 1899.

EASTMAN, H. 1904. New England Ferns and their Common Allies. Boston. 161 pp.
EATON, A. 1836. Manual of Botany for North America. Albany.

-- and WRIGHT, J. 1840. North American Botany; ed. vin.
EATON, D. C. 1879. The Ferns of North America. 2 vols., 1879-1880, 81 plates. Salem,

Mass.
FEE, A. L. A. 1850. Genera Filicum (Cinquieme Memoire sur la Famille des Fougeres).

Exposition des Genres de la Famille des Polypodiace'es. Paris and Strasbourg.
FORD, S. O. [902. The Anatomy of Ceratopteris thalictroides, in Annals of Botany,

16: 95-122, pi. 6. March.
FOWLER. New Brunswick Catalogue.
GILBERT, B. D 1905. Observations on North American Pteridophytes. n, in Fern

Bulletin, 13: 100-104.
GRAY, ASA. 1848-1889. Manual of the Botany of the Northern United States; ed. I,

1848; n, 1856; in, 1858; iv. 1859; v, 1880; vi, 1889.

GOEBEL, K. 1887. Outlines of Classification and Special Morphology, English trans-
lation.

- 1900. Organography of Plants. Transl. by Balfour, 1: 14-17.
GWYNME-VAUGHAN, D. T. 1901. Observations on the Anatomy of Solenostelic Ferns

I. Loxosoma. In Annals of Botany, 15: 71-99-

- 1903. Ditto, II, 1. c., 17: 689-743.

HABERLANDT, G. 1904. Physiologische Pflanzenanatomie, ed. 3.

BIBLIOGRAPHY. 49

HOOKER, W. J. 1840. Flora boreali-americana, 2: 264. London.

- 1846. Species Filicum, vol. i.

- and BAKER, J. G. 1874. Synopsis Filicum, ed. i, 1868; n, 1874.

JANSE, J. M. 1895. Les endophytes radicaux de quelques plantes Javanaises, in Ann.

du Jardin Botanique de Buitenzorg, 14: 53-201, pi. v-xv. (Vol. 14 is dated

1897.)
JEFFREY, E. C. 1897. The Morphology of the Central Cylinder in Vascular Plants, in

Report of the British Assoc. for the Adv. of Sci. 1897, pp. 869-870.
JOHNSON, D. S. 1898. On the development of the leaf and sporocarp in Marsilia

quadrifolia L., in Annals of Botany, 12: 119-147. June.
KELLER, I., and BROWN, S. 1905. Handbook of the Flora of Philadelphia and Vicinity.

Phila. Bot. Club.
KNY. 1875. Die Entvvickelung der Parkeriaceen dargestellt an Ceratopteris thal-

ictroides Brongn., in Nova Acta d. Leopoldinisch Carolinische Akademie der

Naturforscher, 37: 1-80, pi. 18-25.
KUNZE, G. 1848. In Silliman's American Journal of Science and Arts, 2d series, 6:

80-89. "Notes on some ferns of the United Slates," written 1846.

- {850^. Einige Bemerkungen iiber Dicksonia, in Botan. Zeitung, 8: 57-62.

- 1850$. Index Filicum (sensu latissimo), in Linmra, 23: 209-323.

LACHMANN, P. 1877. Structure et croissance de la racine des Fougeres, in Bull. Soc.
bot. Lyon. Not seen.

- 1885. Recherches sur la morphologic et 1'anatomie des Fougeres, in Comptes

Rendus, 101: 603. Not seen.

- 1887. Structure et croissance de la racine des Fougeres, et 1'origine des radicelles.

Bull. Soc. bot. de Lyon.

LAWSON, Geo. 1864. Synopsis of Canadian Ferns and Filicoid Plants, in Trans. Bot.
Soc. Edinb., 8: 20-50. vidi. Also in Edinb. New Philos. Journ., N. S., 19:io2-
116,273-291, 1864; Canacl. Naturalist, N. S., 1: 262-380, 1864. Not seen.

- 1888. The Fern Flora of Canada. A. W. Mackinlay, Halifax. Not seen.
L'HERITIER DE BRUTELLE, C. L. 1788. Sertum Anglicum. Folio, 36 pp., 34 plates.

Not seen.
LINK, H. F. 1841. Filicum species in horto regio botanico Berolinensi culUe. 179 pp.

Not seen.

LINNAEUS, C. 1753. Species Plantarum.

LOWE, E. J. 1867. Ferns: British and Exotic, vol. 8. London.
MACOUN, J. 1890. Catalogue of Canadian Plants. Part v, Acrogens. Geological and

Natural History Survey of Canada, vol. 2. Montreal.

- and BURGESS, F. W. 1884. Canadian Filicinetc, in Trans. Roy. Soc. Canada,

2: 163-227; sec. 4.

MAXON, W. R. 1899. A variety of Dicksonia, in Fern Bulletin 1: 63-64. Not seen.
METTENIUS, G. H. 1856. Filices horti botanici Lipsiensis.
MICHAUX, A. Hort. med. Paris. Cat. not seen.

- 1803. Flora Boreali-Americana, 2: 268.
MOORE, THOS. 1857. Index Filicum.

NAGELI, C., and LEITGEB, H. 1865. In Sitzber. baier. Akad. Wiss.

- 1868. Entstehung und Wachsthum der Wurzeln in Beitr. z. wiss. Botanik, Hft. iv,

pp. 74-160, pis. xi-xxi.

NEWMAN, E. 1841-1851. The Phytologist, 5: 236, fide Moore, 1857.
NUTTALL, THOS. 1818. The Genera of North American Plants, and a catalogue of

the species to the year 1817. Philadelphia.
PARSONS, F. T. 1899. How to Know the Ferns. New York.
POIRET, J. L. M. 1811, Article "Dicksonia," in Lamarck, Encyclopedic Methodique

Botanique. Supplement, tome II, p. 472. Paris, 1811.
PRANTL, K. 1875^. Hymenophyllaceen. Leipzig.

— 1875/5. Verzweigung des Stammes, in Flora, No. 34. Not seen.
PRESL, K. B. 1836. Tentamen pteridographio?, sen genera filicarum pnesertim juxta

venarum decursum et distributionem exposita. Prag.

PROVANCHER, L. 1862. Flore Canadienne. Quebec. 842pp. Not seen.
PURSH, F. 1814. Plants of North America, 2: 664.
RADDI, J. 1825. Plantarum Braziliensium nova genera et species novas vel minus

cognitas. Florentias.
Russow, EDM. 1871. VergleichendeUntersuchungen iiber die Leitbiindelkryptogamen.

in Me"m. de 1'Acad. Imp. de St. Petersbourg, ser. vn, 19: So.
SACHS, J. 1875. Text-book of Botany. English transl.

50 STRUCTURE AND LIFE-HISTORY OF HAY-SCENTED FERN.

SADEBECK, R. 1873. In Verb. Bot. Ver. Prov. Brandenburg, 15.

- 1874. Farnblatt.

- 1878. Vascular Cryptogams in Schenck's Hanclbuch.

- 1898. Pteridophyta in Engler and PrantPs Natiirlichen Pflanzenfamilien, 1-4 A.
SCHKUHR, C. 1809. Vier und zwanzigste Klasse des Linnd'schen Pflanzensystems,

oder Kryptogamische Gewachse.
SCHOUTE, J. C. 1903. Die Stelar-theorie. Jena.

SMALL, J. K. 1903. Flora of the Southeastern United States. New York.
SMITH, j. 1841. Enumeratio Filicum Philippinarum, in Hooker's Journ. of Botany, 3:

392-422.

- 1842. An Arrangement and Definition of the Genera of Ferns, with observations

on the affinities of each genus, in Hooker's London Journ. of Botany, 1: 419-438.

- 1846^, in Bot. Mag. Comp., 36. Not seen.

- 1846^. Catalogue of the Ferns grown at Kew. Not seen.

SPKENGEL, K. 1827. Caroli Linnaei Systema vegetabilium, vol. 4. Not seen.
STRASBURGER, E. 1897. Das Botanische Practicum, ed. v
STRASBURGER, NOLL, etc. 1898. Text Book of Botany, English translation.
SWARTZ, O. 1800. Genera et Species Filicum, in Schrader's Journ. fiir die Botanik,
2: 1-120.

- 1806. Synopsis Filicum.

TANSLEY, A. G., and LULHAM, R. B. J. 1902. On a new type of Fern-stele, and its
probable phylogenetic relations, in Annals of Botany, 16: 157-164. March.

THOMAE, K. 1886. Blattstiele der Fame, in Jahrb. f. wiss. Bot., 17: 129.

TORREY, J. 1843. A Flora of the State of New York. Albany. 2: 502.

UNDERWOOD, L. M. 1888; 1893; 1900^:. Our Native Ferns and their Allies. Ed. i,
1888; vi, 1900.

- 1899. A Review of the Genera of Ferns proposed prior to 1832, in Mem. Torr. Bot.

Club, 6: No. 4.

- 1900*7. Article "DennsUedtia" in Bailey, Cyclopedia of American Horticulture, 1:

472.

— 1900$. Article "Dicksonia," 1. c., p. 480.
VAN TIEGHEM, PH., and DOULIOT, H. 1888. Recherches comparatives sur 1'origine

des membres endogenes dans les plantes vasculaires, in Ann. Sci. Nat. Bot.,

ser. 7, 8: 127-132, 435-438-
VINES, S. H. 1894. Text Hook of Botany.
WATERS, C. E. T903. Ferns. A manual for the northeastern States, with analytical

keys based on the stalks and on the fructification. Holt, New York. 1903.

Cf. pp. 286-290.
WILLDENOW, C. L. 1809. Enumeratio plantarum horti regii botanici Berolinensis.

Berlin.

- 1810. Species plantarum. Ed. TV.
WILLIAMSON, J. 1878. Ferns of Kentucky. Louisville.

WILSON, F. 1897. "Dicksonia pilosiuscula," in Asa Gray Bulletin, 5: 7-9. Not seen.
WOOD, A. 1846. Manual of Botany; editions of 1846 and 1864.

LIST OF ILLUSTRATIONS AND EXPLANATION OF I'LATKS. 51

EXPLANATION OF PLATES.*

Key to index-letters on figures (except Jigs. ^/-

a, initial cell. (/>, inner pericycle. r, root.

c, cortex. ip/i, inner phloem. /•/(,, root-hair.

cj, conjunctive parenchyma. in, inner sclerenchyma. /•/, rootlet.

'e, epidermis. ixt, inner starchy tissue. x, sclerenchyma.

/, fundamental tissue. n, endodertnis. .«/, air-space.

h, hypodermis. p, pericycle. xji, spiral tracheid.

/, rhizogenous cell. j>n, parenchyma. .<</, starchy tissue.

;>, inner fundamental tissue. j>h, phloem. 1 tracheid.

icj, inner conjunctive puren- j>/, plerome. c, sieve-tube.

chyma. I'l'h, protophloem. ./•, xylem.

in, inner endodermis. pjc, protoxylem.

Letters used variously: b, d, m, o, u, h\ 1, 2, 3, etc.

PLATE r.

1. Habitat of D. punctilobula; Massachusetts. Photo by C. E. Waters, Ph.D.

2. Leaves as they grow. From Waters, 1903, by courtesy of the author and publisheis

PLATE 2:

3. Rhizome, natural size, showing fork, leaf bases, and leaf-shoots.

4. Leaf-bud with two unequal leaf-shoots, natural size. 3 and 4 from photo by

C. E. Waters, Ph.D.

5. Portion of pinna showing pinnules, lobes, crenations, sori, and hairs. < about

10. From Waters. 1903, by courtesy of the author and publishers.
PLATE 3:

6-8. Diagrams showing distances between rootlets, natural size.

9. Diagrammatic projection of a piece of stem 5 cm. long, including apex, with

figures to indicate the number of roots attached to each eighth of the

circumference.

10-13. Diagrams of divisions of ro it-cap segment as seen from distal side.
14. Diagrammatic longitudinal section of root-tip, showing the origin of the various

tissues. Walls numbered in order of formation.
15-22. Diagrams of division in segments of root-initial. Walls numbered in order

of formation.
PLATE 4:

20-22. See above.

23. Root-tip in longitudinal section. X 210.

24. Longitudinal section of an anomalous root-cap, showing two layers from one

segment at point marked *. X 210.
25-32. Successive transverse sections of a root-tip. 25, 26 in cap; 27-32 in root; b

indicates root-cap. Initial and second set of segments dotted. All in

the same relative position. X 210.
PLATE 5:

28-32. See above.

33. Transverse section of same root, o. i mm. farther up than 32; />, root-cap. X 210.

34. Rhizogenous cell, undivided, in transverse section of root. X 210.

35. Rhizogenous cell, second division, in similar section. X 210.

PLATE 6:

36. Rhizogenous cell, third division. X 210.

37. First division in rhizogenous cell; longitudinal radial section of root. < 210.

38. Endodermis with rhizogenous cells, tangential view. Reconstruction from

serial longitudinal sections of root. X 210.

39. Ditto: another root. X 210.

40. Oblique tangential view of rhizogenous cell, showing its first three divisions.

X 210.

41. Rootlet initial from longitudinal tangential section of root. X 210.

42. Rootlet initial in transverse section of root; one cap-layer. Cells within the

chain of arrows have arisen by proliferation of cortex. X 210.

43. Ditto; two cap-layers, /;, and digestive layer of cortex, o. X 210.

44. Part of transverse section of nearly mature root. X 210.

45. Transverse section of stele of fully mature root. X 210.

46. Part of transverse section of old root with outer layers shedding off. X too.

* All figures are of ftenntxtdtia pnnctilobula, unless otherwise stated in the description.

52 STRUCTURE AND LIFE-HISTORY OF HAY-SCENTED FERN.

PLATE 7:

47. Root-tip of Aspidium warginale, longitudinal section; a, first periclinal wall;

b, second periclinal; <?, endodermis. X 210.

48. Root-tip of Aspidium molle, longitudinal section; e, endodermis. X 210.

49. Root-tip of Diaymochlcena lunulata, longitudinal section; e, endodermis. X 210.

50. Root-tip of Ceratopteris thalictroides, longitudinal section. X 210.

51. Root-tip of Onoclea sensibilis, longitudinal section; e, endodermis. X2io.

52. Cortex of root of Dennscrtdtia punctilobitla, showing mycorhiza, o\ transverse

section of root. X 210.

53. Ditto. Longitudinal section. Entrance of fungus through a root-hair. ( 210.

54. Rootlet, transverse section. Pericycle dotted ; sextant walls heavy. X 210.

55. Junction of pericycle of root and rootlet. X 210.

56. Junction of xylem of root and rootlet; longitudinal radial section of root.

X 200.

57. Ditto; tangential to rootlet. X 200.

PLATE 8:

58. Two tracheids from the junction of the vascular bundles of root and stem.

Macerated and teased out. X 210.

59. Young root in longitudinal section of stem apex. From point marked 3 in fig.

70. X 210.

60. Longitudinal section of older root, about to pass out from stem; />, root-cap.

From point marked 4 in fig. 70. C 210.

61. Junction of root shown in fig. 60 with stem bundle. X 210.

(>2. Sharply bent tracheid at junction of root and stem, from same section as fig. 99.

X 210.
PLATE 9:

63. Part of transverse section of stem. Area included in dotted lines at 2 in

fig- 97-

64. Part of longitudinal section of stem. Area included in dotted line u in fig. 70.

65. Ends of sieve-tubes of stem; macerated and teased.

66. Diagram of node; longitudinal section of stem and leaf-base; 6, vascular bundle.

PLATE 10:

67. Rhizome, transverse section. Photomicrograph.

68. Vascular bundle of petiole, transverse section. Photomicrograph.

69. Rachis of leaf, transverse section. Photomicrograph.

PLATE n:

70. Diagram of stem apex, longitudinal section, i, 9, dotted outlines of leaf rudi-

ments, showing their position relative to the stem apex. 2, 3, 4, 7,8,
location of root-tips of various ages (cf. figs. 59, 60 1. 5, beginning of
protophloem. 6, point where endodermis and pericycle are separated.
10, beginning of lignification in xylem. u, area drawn in fig. 64.

71. Stem-initial in transverse section. X 210.

72-75. Segmentation and sectioning in the apex shown in fig. 71. Segments num-
bered in order of age.

76. Isolated cells of sclerotic medulla; macerated and teased.

77. Ditto.

78. End of scalariform tracheid of stem; macerated and teased.
7<j. Ditto.

80. Isolated cells of sclerotic cortex of stem; macerated and teased.

8 1. Isolated cells of conjunctive parenchyma of stem; macerated and teased.

82. Vascular bundle of stem and leaf, with leaf-shoot dissected out and viewed from

the side; o, leaf-gap; m, leaf-shoot; /;-, leaf-trace; //, anterior end of por-
tion of vascular bundle of stem. Drawn from nature by Miss M. E.
Rogers.
PLATE 12:

83-87. Diagrams of successive cross-sections of a fork of a stem, with ventral leaf.
Dotted line indicates boundary between sclerotic and starchy tissues;
vascular bundle section-lined; /r, leaf-trace.

88-92. Diagrams of successive cross-sections of rachis, showing departure of rib of
pinna. Dotted line bounds sclerenchyma; solid line is endodermis;
xylem is section-lined; b, trace of pinna. Fig. 88 is lowermost, 92 upper-
most.

LIST OF ILLUSTRATIONS AND EXPLANATION OK PLATKS. 53

PLATE 12, continued:

93-96. Diagrams of successive cross-sections of petiolar bundle giving off a leaf-
shoot bundle. Conventional signs as in figs. 83-87; ;//, leaf-shoot.

97. Diagrammatic cross-section of leaf-shoot, near its origin. Conventional signs as

in iigs. 88-92; 2 indicates part drawn in fig. 63.

98. Isolated tracheids of fork of stem.

99. Diagram of fork of stem, longitudinal section. Conventional signs as in figs.

83-87.

100. Diagrammatic cross-section of node; stem to right. Conventional signs as in

figs. 83-87.

101. Diagrammatic transverse section of base of petiole. Conventional signs as in

figs. 88-93; dt region of stomata.

102. Diagram of junction of root and stem, longitudinal section. Cortex and

medulla section-lined.
PLATE 13:

103. Stem-initial with four segments, in transverse section, with leaf-rudiment, d, at

left. X 210.

104. Transverse section of stoma. After Copeland, 1902.

105. Glandular hair of leaf, with two gland-cells. X 210.

106. Longitudinal section of stem apex; heavy lines at right show boundary of

vascular bundle; d, intercalary growth of hair; //, origin of hair.
Composite. X 210.

107. Protophloem in transverse section of stem near apex. C 210.

108. Transverse section of stem-bundle at margin of leaf-gap, showing continuity of

inner and outer vascular tissues. X 210.

109. Oblique transverse section of stem-initial, with leaf-rudiment (b) in fourth

segment. X 360.
no. Protophloem and earliest lignified xylem in transverse section of stem near

apex. See lettering on fig. 107.
PLATE 14:

in. Lower epidermis of leaf. X 210.

112. Upper epidermis of leaf. C 210.

113. Transverse section of leaf; chloroplasts diagrammatic. ( 210.

114. Diagrammatic transverse section of rachis near summit; xylem shaded; ///,

parenchymatous region connecting with stomata; tr, trace of pinna.

115. Inner surface of indusium, near margin. <C 210.

116. Cavity parenchyma (d] in transverse section of rachis. < 210.

117. Ditto, in longitudinal section. X 210.

118. Early stage of leaf-initial in transverse section of stem apex.

119. Outer surface of indusium. X 210.

120. Inner surface of indusium, near base, with stomata. ( 210.

121. Formation of leaf-initial in stem apex; later stage than fig. 118. Transverse

section.
PLATE 15:

122-127. Serial transverse sections of leaf-rudiment, showing transition from four-
sided to two-sided initial; 122 is through apex of rudiment, 127 through
its base.

128, 129. Longitudinal sections of stem apex with leaf-rudiment, taken 0.025 mm.
apart; w, region of stem-initial; a, leaf-initial; b, stem-initial.

130. Longitudinal section of leaf-rudiment in radial longitudinal section of stem.

X 360.

131. Transverse section of two-sided leaf-initial, from a leaf with three pairs of

pinnae.

132-134. Serial sections (obliquely sagittal) through one end of leaf-initial, showing
segments and sections. From a leaf with four pairs of pinnae.

135. Longitudinal section of leaf-initial; from a leaf with three pairs of pinna.'.

136. Longitudinal section of leaf-tip; no pinna;.

137. Surface section of leaf-tip with marginal initials.

PLATE 16:

138. Longitudinal section of leaf with five pairs of pinnae, showing loss of initial.

Segments i and 2 formed successively on same side of apical cell.
X 360.

139. Sagittal section of leaf-rudiment with three pairs of pinnae, becoming circinate.

Heavy lines bound segments.

54 STRUCTURE AND LIFE-HISTORY OF HAY-SCENTED FERN.

PLATE 16, continued:

140. Transverse section of rachis near apex of a leaf with three pairs of pinna.' and

a single initial cell, showing sectioning of marginal cells, m.

141. Ditto; leaf with seventeen pairs of pinna?, growing by a group of marginal

initials.

142-143. Transverse section of rachis of a leaf with seven pairs of pinna:-, showing
cessation of division in marginal cells, m. Fig. 142 is between second
and third pairs of pinnae. < 360.

144. Sagittal section of apex of a leaf with nine pairs of pinna-, and growing by a

group of marginal initials, ;//. X 360.

145. Horizontal section of tip of pinna; »i, growing point. From a leaf with eleven

pairs of pinnae.

146. Transverse section of pinna near apex; ///, marginal cell.

147. Transverse section of leaf through a developing pinna; /;/, marginal cell. From

same leaf as fig. 146.
PLATE 17:

148. Horizontal section of developing pinnule; lobes and sinus.

141;. Horizontal section of teeth of pinnule lobe, with developing veinlets (shaded).

150. Dorsiventral section of developing lamina.

151. Transverse section of leaf with rudiment of sorus on margin; m, mother-cell of

first sporangium; u, indusium. < 210.

152. Outline of pinnule of unfolding leaf. X 42.

153. Surface of pinnule shown in fig. 152: »i, rudiment of stoma.

154. Longitudinal section of young sporangium; central cell just formed.

155. Transverse section of leaf-margin through a mature sorus; d, placenta; //, indu-

sium. X 210.

156. Oblique longitudinal section of developing sporangium with one central cell.

157. Longitudinal section of young sporangium; stalk and wall segments cut off;

cap not yet formed. : 360.

158. Sagittal section of rudiment of sorus; ft, indusium; i, 2, successive sporangia.

X 360.

159-167. Sections of developing sporangia, showing stages as follows: Fig. 159,
first cleavage in mother-cell; d, placenta. Fig. 160, three-celled rudi-
ment. Fig. 161, first tapetal cell. Fig. 162, first tapetal layer com-
plete (on right), leaving the archesporial cell. A three-celled rudiment
at left. Fig. 163, division of the tapetal layer. Fig. 164, four arche-
sporial cells in equatorial-plate stage, dividing to make eight. Figs.
165, 166, adjacent sections of two-celled archesporium, dividing into
four. Fig. 167, spore mother-cells; tapetum degenerating.
PLATE 18:

161-167. See above.

168. Tetrads, with fragment of tapetum.

169. Paraphysis arising from placenta.

170. Sagittal section of sporangium, showing the spore mother -cells just before

synapsis.

171-172. Mature sporangia, from opposite sides.
173. Mature paraphysis. : 360.
174-175. Surface views of spores.

176. Transverse section of stalk of sporangium.

177. Germinating spore. : 360.

178. Three-celled protonema, short type.

179. Two-celled protonema. < 360.

i So. Three-celled protonema, medium length.

181. Four-celled protonema.

182. Protonema with short basal cells. X 360.

183. Six-celled protonema.

184. Three-celled protonema. Long type.

185. Five-celled protonema with two-sided initial. < 360.

1 86. Base of a prothallus without protonema.

187. Protonema with two segments from initial. 360.
iSS. Five-celled prothallus without protonema. < 360.

LIST OF ILLUSTRATIONS AND EXPLANATION OF I'LATKS. 55

PLATE 19:

189. First longitudinal division in protonema.

190. Seven-celled protonema with two-sided initial. X 360.

191. Side view of tig. 190.

192. Six-celled protonema; beginning of two-sided initial.

193. Same as fig. 192, showing length of rhizoids.

194. Apical growth established; one antheridium.

195. More advanced apical growth, drawn from a specimen plasmoljzed in salt

solution in order to show the walls.

196. Long, six-celled protonema with two-sided initial just established.

197. Protonema with irregular apex.

198. Dwarf male prothallus. Westport, Maryland, September 25, 1905. Four

prothallial cells, three antheridia.

199. Normal male, becoming cordate.

200. Group of marginal initials just established. X 210.

PLATE 20:

201. Papillar outgrowth on dorsal surface of female prothallus. X 210.

202. Two-celled archegonial rudiment, d, in sagittal section of prothallus. X 360.

203. Upper surface of prothallus; single initial giving place to a group. X 210.

204. Apical growth with a group of initials; horizontal section. : 360.

205. Vertical transverse section of prothallus, 0.05 mm. back of notch, /. f., about

the line i-i in fig. 204.

206. Three-celled archegonial rudiment, d, in sagittal section of prothallus, cut along

the line 2-2 in fig. 204. X 360.

207. Three-celled archegonial rudiment, d; first cleavage in neck-rudiment. < 360.

208. Archegonial rudiment; neck and central cell. X 360.

209. Cutting off neck canal-cell. X 360.

210. Longitudinal section showing cleavage in neck cell. X 360.

211. Longitudinal section of young archegonium showing neck bending over. X 360.

212. Surface view of neck of nearly mature archegonium. X 360.

213. Ditto; young archegonium. X 360.

214. Archegonial rudiment, d, one cell. X 360.

215. Longitudinal section of prothallus, showing neck canal cell with two nuclei.

' 360.

216. Formation of ventral canal-cell. X 360.

217. Axial cells of archegonium complete; neck immature. : 360.

218. Neck canal-cell with three nuclei.

219. Mature archegonium; stained with iron hiematoxylin. Ventral wall complete.

X 360.

220-222. Transverse sections of mature neck at summit, middle, and base, respec-
tively. X 360.

223. Egg-cell ready for fecundation.

224. Fertilized egg in longitudinal section of archegonium. Unsuccessful sperms in

mucilage of neck. X 360.
PLATE 21:

225. Longitudinal section of antheridial rudiment, one-celled. X 360.

226. Ditto, two-celled. : 360.

227. Surface view of a lobe of an old male prothallus. First division in central cell

of antheridium above. X 210.

228. Ditto; three-celled antheridium.

229. Side view of antheridium with long basal cell.

230. Antheridium nearly mature; vertical section. X 360.

231. Rhizoids of female prothallus. X 210.

232. Optical section of three-celled antheridium; cleared in glycerin. : 360.
233,234. Octants of embryo. Transverse sections of prothallus, 233 being 15 n

anterior to 234; />, stem-octant; d, rudiment of first leaf; r, root-quad-
rant; ?/, foot.

235, 236. Sagittal sections of embryo and calyptra, 25 M apart; b, stem initial;^/, mar-
ginal cell of first leaf; u, foot.

237. Longitudinal section of old antheridium.

56 STRUCTURE AND LIFE-HISTORY OF HAY-SCENTED FERN.

PLATE 22:

238. Transverse section of protostele below first leaf-gap. X 210.

239. Transverse section ot stem through the first leaf-gap, o. No inner phloem.

Siphonostelic structure occurs o.i mm. higher up; /r, region of first leaf-
trace. X 210.

240. Transverse section of siphonostele between first and second leaves.

2\\. Transverse section of siphonostele midway between third and fourth leaves.

242-245. Serial transverse sections of the center of the stele at the origin of the
inner endodermis. 241 to 242 is 60 ju; 242 to 243 is 10 /j.] 243 to 244 is
20 /u.; 244 to 545 is 70 /x. All between third and fourth leaves. X 210.
PLATE 23:

246. Sagittal section of young fern attached to prothallus; m, stem-initial; r>, calyptra

r, root; tr, first leaf; ?/, prothallus. X 75.

247. Detail of transition from root to stem, from same series as fig. 246; ;;/, anterior;

r, posterior; //, upper part.

248. Slightly oblique transverse section of petiole of third leaf of sporeling; d, upper

side. X2io.

249. Glandular hair from leaf of adult plant. < 43.

250. Transverse section of petiole of first leaf, showing sectioning of marginal cells.

Endodermis dotted; ;//, marginal cell. X 360.

251. Sinus of leaf-margin of seedling; dotted cells are epidermal. Surface view of

cleared specimen. X 350.

252. Young rhizoids and tips of mature ones from root of three-leaved sporeling of

fig. 267. X 350.

253. Optical section of spongy parenchyma of first leaf of sporeling; cleared in gly-

cerine. X 350.

254. Epidermis and developing stomata on sporeling leaf; w, stoma mother-cells.

X43-
PLATE 24:

255. Horizontal section of calyptra (//) and embryo through root and rudiment of

first leaf (tr). X 360.

256. Diagrammatic cross-section of sporeling plant through third leaf-gap; tr, petiole

0, leaf-gap.

257. Diagrammatic cross-section of stele and starchy cortex of sporeling stem at

fourth leaf-gap; o, leaf-gap. X 75.

258. Hairs of third leaf of sporeling No. 3; />, acicular; »i, moniliform ; u, glandular.

X 350.

259. First leaf of sporeling No. 2. X 5.

260. Second leaf of same plant. X 5.

261. Second leaf of plant No. 3. X 5.

262. Third leaf of plant No. 3. X 5.

263. Third leaf of another plant. C 5.

264. Fourth leaf of plant No. 3. C 5.

265. Fourth leaf of plant No. 4. : 5.

PLATE 25:

266. Lower epidermis of first leaf of sporeling. X 350.

267. Three-leaved sporeling (No. i) with portion of prothallus (it) still attached.

1, 2, 3, first, second, and third leaves.

268. Surface view of petiole of second leaf of sporeling (No. i), with stoma.

269. Sporeling stem with roots and leaves cut off. Drawn from nature by Miss M.

E. Rogers.

270. Forked stem of seedling, i, primary root leading up to the original simple

stem. Drawn from nature by Miss M. E. Rogers.

Plate 1

l^aJSEESPW J 41&. !

:, 3& ^|r -:^Pt

FIG. i. Habitat of D. punctilobula, Massachusetts.

FIG. 2. Leaves as they grow.

Plate 2

FIG. 3. Rhizome, natural size, showing fork, leaf-bases, and leaf-shoots.

FIG. 4. Leaf-bud with two unequal leaf-shoots, natural size.

FIG. 5. Portion of pinna showing pinnules, lobes, crenations, sori, and hairs. X about 10.

6

u

8

10

\

Vrnt

\\\ \\\ \v\

II

12

¥1 fl X K

\zmviivvi K x H iv

PLATE 4.

20

PLATE 5.

PLATE 6.

PLATE?.

47

PLATE;

PLATE 9.

Plate 10

us?

68

69

":^§^^M^;;:^

PHOTOMICROGRAPHS.
FIG. 67. Rhizome, transverse section.
FIG. 68. Vascular bundle of petiole, transverse section.
FIG. 69. Rachis of leaf, transverse section.

PLATE 11.

PLATE 12.

. iph .
in 1S is || phst s

PLATE 13.

PLATE 14.

in

IIQ

120

121

PLATE 15.

PLATE 16.

PLATE 17.

PLATE 18.

PLATE 19.

PLATE 20.

PLATE 21.

233

236

PLATE 22.

PLATE 23.

PLATE 24.

255

PLATE 25.

266

3V 30

M