73,HS3Q9 BOSeO-9

W. G. FARLOW

HARVARD BOTANICAL MEMOIRS

es

Avr a meeting of the Botanical Department held
Oct. 20, 1903, the following vote was passed:

““Under the head of Harvard Botanical Memoirs it is
proposed to include all quarto publications issuing from
the Gray Herbarium, the Cryptogamic Herbarium, and
the Botanical Laboratories of Harvard University, including
theses presented for the degrees of Ph.D. and S.D. in Botany.
Inasmuch as some of the future publications are likely to be
continuations of subjects treated in quarto papers already
published, it seemed desirable to begin the numbering of the
Memoirs with the year 1880, the date of the first quarto
publication of any member of the botanical staff at present
connected with Harvard University.”

At a meeting on Nov. 25, 1916, it was voted to dis-
continue the series of Botanical Memoirs. In all, nine
numbers have been issued, the titles of which are given
below.

I. The Gymnosporangia or Cedar-Apples of the United States. ,
By W. G. Farlow. Anniversary Memoirs, Boston Soc.
Nat. Hist. 1880. Pp. 38. Pls. 1 and 2.

II. The Entomopthoreae of the United States. By Roland /
Thaxter. Mem. Boston Soc. Nat. Hist., IV, No. 6. Pp.
133-201. Pls. 14-21. April, 1888.

III. The Flora of the Kurile Islands. By K. Miyabe. Mem.
Boston Soc. Nat. Hist., IV, No. 7. Pp. 203-275. Pl. 22.
Feb. 1890.

IV. A North American Anthurus: its Structure and Development.
By Edward A. Burt. Mem. Boston Soc. Nat. Hist., III,
No. 14. Pp. 487-505. Pls. 49 and 50. Oct. 1894.

V. Contribution towards a Monograph of the Laboulbeniaceae. _/
By Roland Thaxter. Mem. American Acad. of Arts and
Sci. Boston. XII, No. 3. Pp. 189-429. Pls. 1-26.
Presented May 8, 1895. Issued Oct. 14, 1896.

VI. The Development, Structure, and Affinities of the Genus
Equisetum. By Edward C. Jeffrey. Mem. Boston Soc.
Nat. Hist., V. No. 5. Pp. 155-190. Pls. 26-30. April,
1899.

“nt

VII. The Comparative Anatomy and Phyllogeny of the Coni- 4
ferales, Part I. The Genus Sequoia. By Edward C.
Jeffrey. Mem. Boston Soc. Nat. Hist., V, No. 10. Pp.
441-459. Pls. 68-71. Novy. 1903.
VIII. The Comparative Anatomy and Phyllogeny of the Coni-
ferales, Part II. The Abietineae. By Edward C. Jeffrey.
Mem. Boston Soc. Nat. Hist., V1, No. 1. Pp. 1-37. Pls.
1-7. Jan. 1905.
IX. Contributions towards a Monograph of the Laboulbeniaceae, |/
Part II. By Roland Thaxter. Mem. American Acad. of
Arts and Sci., XIII, No. 6. Pp. 219-469. Pls. 28-71.
June, 1908.

J o\? Ys 2

a , ‘ 1% , id
az ~ “ig! 4
a) Pee . .
‘* aed Ay Rae ee 4
‘ , €
vb ae ers ,
‘ . Lf Art .
’ rar t ‘ : ‘
*, 4 t ‘ 4 .
be erer Fh 4 ’ 4 )

g 4 ig? : - ,
OF iar We ea Re

Wiens
tees

=
=
<
=

‘
’

| SOCIETY OF NATURAL HISTORY

OF THE

~

VOLUME 6, NUMBER 1.

AND PHYLOGENY OF THE CONIFERALES

PART 2.—THE ABIETINEAE,

’

Sodan
r

JANUARY, 1905.

TH SEVEN PLATES.
BOSTON
_ PUBLISHED BY THE SOCIETY.

WI" “Fs

By EDWARD ©. JEFFREY.

4

NAT

1. THe CompaARATIVE ANATOMY AND PHYLOGENY OF THE CONIFERALES.

Part 2.— THE ABIETINEAE.!

By Epwarp C. JEFFREY.

(Read November 2, 1904.)
INTRODUCTION.

In the first number of this series the writer (:03) has called attention to
certain features of the anatomy of the genus Sequoia, which seemed to indicate for
that genus an Abietineous origin. The results obtained in the case of Sequoia made
it desirable that the Abietineae should be somewhat fully examined as a preliminary
to the study of the other orders of the Coniferales. This procedure seems further
advantageous, because the Abietineae are perhaps the most clearly defined and most
accessible of existing Coniferous orders and are at the same time a geologically very
ancient group.

The importance attachable to the anatomy of the Gymnosperms has recently
been greatly imcreased by the discovery of the seeds of certain of the Cycadofilices.
This alliance established by Potonie ('99, p. 160) on the basis of the morphology
and anatomy of the vegetative organs alone, no longer ago than 1899, has recently
received the fullest confirmation from the epoch-making discoveries of Oliver and
Scott (:03) and Kidston (:03), which, as has been stated by Zeiller (:04), are des-
tined to make a revolution in our conception of the nature of the Palaeozoic flora.
By these discoveries, especially when taken together with older, similar ones in the
case of the Calamites and Sigillarians, the taxonomic value of the anatomy of plants
is placed at once on as secure a footing as that of animals since the time of Cuvier.
The study of anatomy, which is so essential and which has recently yielded such impor-
tant results in the case of fossil plants, has been much neglected in those now living,
largely on account of the absence of proper perspective, which can alone be supplied
by the knowledge of the older forms. In the case of the Angiosperms this per-
spective is still unfortunately lacking by reason of our entire ignorance of the
nature and structure of their ancestry; but for the Gymnosperms the anatomical
researches of the British, French, and German palaeobotanists have now reached a

1 Harvard Botanical Memoirs. — No. 8.

2 EDWARD C. JEFFREY ON

fullness which makes them of the greatest value in the interpretation of the phylo-
genetic relationships of the existing Gymnospermous groups. Extremely important
results have been obtained in recent years by the anatomical method, in the case
of the Ginkgoales and Cycadales; but the Coniferales can scarcely be considered
to have been properly investigated in the light of recent additions to our knowl
edge of the structure of the Cycadofilices, Cordaitales, and other ancient and tran-

sitional groups. This statement is made because investigations of recent date on

o
g
the Coniferales either do not concern themselves with a consideration of the fossils
at all, or are devoted to some special theme, such as the structure of the wood,
the anatomy of the leaf, ete. Modern work on the anatomy of plant fossils empha-
sizes the fact that all the organs and tissues should be examined, where successful
comparisons are to be made. It will accordingly be the aim of the present investi-
gation to include all the structural features of the Abietineae, so far as they seem

to have any bearing on the phylogeny of the order.

PSEUDOTSUGA.

In a transverse section of a vigorous yearling shoot of Pseudotsuga douglasii
Carr., one sees the comparatively thick fibrovascular ring surrounded by a_ thick
cortex. The wood of this genus is characterized by the presence of resin canals
of which there are generally present about three rows. Occasionally one finds shoots
in which the resin ducts of the wood are quite absent. The pith is irregular in
outline, on account of the occurrence of projections from its periphery correspond-
ing to the outgoing leaf traces. These rays of medullary parenchyma become deeper
and deeper as the leaf trace passes outwards from the medullary crown towards
the cortex, and where the trace is about to leave the fibrovascular rmg there is a
radial stripe of parenchyma several cells broad extending entirely through the woody
cylinder. This is the foliar gap. Very shortly after leaving the medullary crown,
the leaf trace becomes divided into two by a radial row of parenchyma cells, which
as the trace begins to approach the outside of the woody cylinder, becomes two and
even three layers of cells wide. One or both of the leaf traces at this time may
contain a resin canal, which disappears as soon as the trace begins to pass into the
phloem. Figure 4, plate 1, shows the topography of the passage of the foliar trace
out of the central cylinder and demonstrates that it is distinctly double. It is
important to note this point because Bertrand (’74) states that Pseudotsuga, together
with Tsuga and Larix, is characterized by a single undivided leaf trace. In figure 5,
plate 1, the leaf trace is seen at some distance out in the cortex, and is obviously

ye er

ANATOMY AND PHYLOGENY OF THE CONIFERALES. 3

even more strikingly double than when it is leaving the central cylinder. The
foliar trace continues as geminate strands for a short distance into the base of the
leaf; but soon the two fibrovascular bundles become more or less intimately fused.
In a section at some height in the leaf, the trace loses all appearance of being
a double strand. This is clearly shown in figure 3, plate 1, which is from a section
taken at about the region of the lower third of the leaf of Psewdotsuga douglasii Carr.
It will thus be seen that the leaf trace of this species becomes double at an early
stage in its progress towards the leaf, and that its double nature only becomes
obscured after having penetrated some distance into the lamina. The importance
of this observation will be apparent at a later stage.

The root of Pseudotsuga douglasii is generally diarchous, although examples have
been found in which there were as many as three clusters of primary xylem. The
primary metaxylem contains no resin canal; but just outside each of the two clusters
of protoxylem is a single duct. ‘These features can be seen in figure 2, plate 1. The
two first formed resin canals are lined with thin walled, glandular cells, such as occur
in the resin ducts of the cortex in the Coniferales in general, and in the ligneous resin
ducts of Pinus. The lining of the canals is further particularized by the fact that the
cells composing it are quite unlignified and stain a dark blue with haematoxylin. In
this respect the tissue in question presents a marked contrast to that lining the later
formed ducts of the secondary wood in the root and shoot of the species under discus-
sion, which in most cases is thick walled and distinctly lignified. It not infrequently
happens that the primary radical resin canals of Pseudotsuga communicate by radial
anastomoses with the resin canals of the cortex. The walls of the radial connecting
canals in this case are bounded by the same thin walled, resin secreting cells as are found
in the ducts which they unite. These radial anastomoses are most frequent in the
neighborhood of an outgoing rootlet, but are by no means confined to this position.
In figure 2, plate 1, on the left, may be seen one of the radial canals in question. The
plane of section does not include the cortical canal into which it opens. In the cortex
of both root and shoot of Pseudotsuga are sclerotic cells; but these do not occur in the
phloem as in T’suga and Abies.

In figure 1, plate 1, is to be seen a section through a wounded portion of the wood
in Pseudotsuga. On the left may be seen the margin of the layer of wound tissue,
which is closing over the surface of the imjured wood. By following this layer
towards the right we reach numerous large resin canals along its inner border. These
are traumatic resin ducts such as occur almost invariably in the injured wood of the
Abietineae. They are characterized by their number, by the fact that they occur in
close tangential rows, the ducts of which generally intercommunicate tangentially, and

4 EDWARD C. JEFFREY ON

by the nature of the cells lining their walls. They are apparently for the purpose of
sterilizing the wound by pourmg large quantities of resin over its surface. Their
occurrence in close tangential rows with tangential intercommunication of the ducts
is a source of mechanical weakness to the wood, which it is not surprising to find to
some extent overcome by the nature of the cells lining the canals. The resiniparous
canals in the case of traumatic resin ducts are lined by thick walled, resin secreting cells,
which are able to accomplish their function of pouring out a resinous secretion by the
presence of numerous simple pits.

In the wood of the cone of Psewdotsuga douglasii Carr., there is a single row of
resin canals near the medulla. In longitudinal sections the tracheids of the wood
in the female reproductive axis are without the spiral bands, which are a well known
characteristic of those of the vegetative axis (Penhallow, 96). Although tracheids with
tertiary spiral thickenings are absent in both the axis and the scales of the cone in this
species, they occur in the leaf traces. This would seem to lead to the inference that their
absence in the case of the female reproductive axis is rather due to the thickness of
the walls of the tracheids in this organ than that the spiral thickenings in question are
not a primitive feature of Pseudotsuga. This inference seems strengthened by the
fact that the spiral markings are sometimes absent in the case of the thick walled, aes-
tival tracheids of the vegetative axis. It does not follow, however, that the occurrence
of spirally marked tracheids in the wood of Pseudotsuga is any indication of close affinity
with the Taxineae, in which they also oceur, since the same phenomenon is sometimes
found in the hard pines and in Larix (Penhallow, :03).

LARIX.

A transverse section of the yearling stem of Larix americana Michx. presents very
much the same appearance as one of Pseudotsuga. There is in this genus, however, an
absence of the sclerenchymatous elements found in the pith and cortex of the last
described genus. The leaf bases, too, are more crowded and more prominent and present
a rounded contour. There is as a rule but a single layer of ligneous resin canals and this
is present in the summer wood. The root of Larix americana somewhat resembles that
of the genus described above, and has the same commissural radial canals connecting the
resin ducts, subtending the two clusters of protoxylem, with those which occur in the
so called secondary cortex. This feature is shown in figure 6, plate 1. The lining of
the resin canals facing the angles of the primary wood is of the same thin walled, resi-
niparous cells, which have been described above for Pseudotsuga. The subsequently
formed resin ducts of the wood of the root have this thin walled, glandular epithelium

ANATOMY AND PHYLOGENY OF THE CONIFERALES. 5

more or less completely replaced by thick walled, resiniparous cells, which are at least
partially lignified. There are, however, nearly always some of the more primitive, thin
walled cells remaining. The radial canals connecting those opposite the primary wood
with similar canals in the secondary cortex have their inner surface also lined with the
thin walled type of gland cells. In the stem of Larix, as has been pointed out by Pen-
hallow (’96, :03), the resin ducts of the wood are coated with thick walled, resiniparous
cells, presenting in this respect a marked contrast to Pinus.

The leaf trace of Larix is described by Bertrand (’74) as resembling that of Tsuga
and Pseudotsuga in remaining a simple undivided fibrovascular strand. An examination
of thin sections through the yearling stem of Larix americana has shown that the foliar
trace here is also double, as in Pseudotsuga described above. The trace sometimes shows
indications of duplication even before it leaves the central cylinder of the axis, but gen-
erally does not become double until it has advanced some distance into the cortex, in its
upward course into the leaf. Figure 7, plate 1, shows the condition of the leaf trace at
about the level at which the resin canals of the mesophyll have made their appearance.
As is well known (Strasburger, ’71), the resin ducts of the fundamental tissues of the leaf
in the Abietineae are not generally continuous with the similar canals of the cortex of the
axis. At this height the leaf trace as seen in transverse section is obviously double, the
tracheary strands bemg separated from each other by two or three rows of parenchyma
cells. These two fibrovascular bundles become intimately fused in the base of the free
portion of the leaf, and it is no longer possible to distinguish any signs of duplication.

PICEA.

In Picea nigra Link, a section of the yearling stem shows the presence of abundant
sclerification in the pith. Sclerenchymatous strands are also present in the cortex, sub-
tending the outgoing leaf traces. The woody cylinder normally contains but a single
row of resin canals which are nearly midway between the beginning and the end of the
annual ring. These canals differ somewhat from those of Larix in their lining, for their
parietal glandular tissue is partially thick walled and partially made up of thin walled
elements. The latter are always present to some degree and no doubt are responsible
for the formation of the tyloses which occur in the canals of Picea, in contrast to those
of Larix, where both tyloses and thin walled elements are for the most part absent.
The leaf bases of Picea are sharp pointed instead of rounded in transverse section, as is
the case in Larix. They are further particularized by the presence of two small resin
canals, which in the mature stem end blindly upwardly, although they are inferiorly in
communication with the much larger resin canals of the cortex of the axis. As Mayr

6 EDWARD C. JEFFREY ON

(84) has pointed out, Picea in contrast to the other Abietineous genera, has the resin
canals of the leaves at first continuous with those of the stem, but the free communication
between the two systems is cut off at an early date by the formation of absciss periderm.
This continuity between the foliar and the axial systems of resin canals appears to have
been the primitive condition for the Abietineae, but as has been indicated above, has
become obsolete in most of the existing species.

Figure 8, plate 1, shows the effect of jury on the structure of the wood in Picea
mgra. The sparse and small resin canals present in the normal wood are replaced by
close tangential rows of resin ducts of much larger lumen, in the folds of traumatic wood,
which in the course of time grow over the injured surface of the stem. The reaction
which leads to the formation of the tangential rows of canals in the case of injury, is of
course much more intense in the immediate vicinity of the wound, and fades away tan-
gentially in the newly formed wood. The traumatic resin canals in question communi-
cate tangentially and have doubtless the function of pouring out resin over the wounded
portion of the stem. It is an interesting fact, that the formation of these tangential rows
of traumatic resin canals in the Abietineae is most marked in those genera in which the
resin canals are normally somewhat sparse or entirely absent. In Pinus, where the resin
ducts of the wood are abundant in both the vertical and the horizontal planes, a marked
reaction leading to the formation of tangential rows of traumatic resin ducts seldom or
never occurs. In Picea, Larix, and Pseudotsuga, on the other hand, their formation is an
almost invariable sequel to injury, and this is even more strikingly the case in Abies,
Tsuga, Cedrus, and Pseudolarix, which are ordinarily described as entirely lacking ligne-
ous resin ducts. In the genus at present under discussion, I have noticed a strong forma-
tion of traumatic resin canals, not only as the result of mechanical injury, but also as
a sequel of the attacks of Chermes abietis, and of parasitic rust fungi producing witches’
brooms.

In Picea the leaf trace is described by Bertrand (’74) as being a double one by the
presence of a distinct median medullary ray. In P. nigra the foliar trace leaves the
central cylinder of the axis as a single strand of very small size. As it reaches prox-
imity to the base of the leaf it becomes distinctly double, as may be seen in figure 9,
plate 2. In P. engelmanm, which was also examined in this connection, the trace is
somewhat larger and the duplication correspondingly more marked. The root in Picea
does not differ in any marked features from those of Larix and Pseudotsuga.

ABIES.

In Abies balsamea Mill., resin canals are normally quite absent in the wood of the
shoot. This feature is illustrated by the central portion of figure 10, plate 1. In

ANATOMY AND PHYLOGENY OF THE CONIFERALES. 7

the two last annual rings of the figure, dense tangential rows of resin canals can be
clearly seen on either side of an injured part of the stem. Figure 11, plate 2, shows
a portion of the traumatic wood on the left of figure 10 more highly magnified. In this
case it is easy to make out that the resin canals are more or less strikingly in tangential
communication with each other, and that they occur in the spring wood of each of the
annual rings formed subsequently to the injury of the stem. The canals in question
pour out a large amount of resinous secretion over the surface of the injured wood, and
by their presence as a reaction to injury only afford a much more economical provision
for antisepsis than the numerous resin canals which occur throughout the wood of Pinus.
In the annual rings of the stem shown in figure 11, both in those containing resin canals
and in those without resin canals, may be seen certain minute dark dots. These are
especially noticeable, in spite of the low magnification, in the annual ring immediately pre-
ceding those which contain resin canals. These dots are the so called resin cells, which
are frequently characteristic of those Coniferous woods which do not contain resin canals
under normal conditions. The dark color which is a feature of the cells in question is
due to the presence of a tanniniferous substance. In figure 12, plate 2, is seen a part
of the woody tissue from the axis of a witches’ broom of Abies balsamea, It is clear
from this figure that resin canals of the tangentially crowded traumatic type are also
present as the result of the attack of the fungus producing the witches’ broom. This
feature has been described by Anderson (’97) and other pathologists. It is to be further
noted that the place of occurrence of the rows of traumatic resin ducts is in general the
same in the species under consideration, viz., in the spring wood.

In figure 13, plate 2, is seen part of a section of an injured stem of Abies magnifica
Murr. The area of the figure is all contained within the breadth of a single annual ring,
yet there are four rows of traumatic resin canals. This is the largest number occurring
within a single annual ring of wood which has come under my notice. Generally in the
various species of Abies there is but a single row of such ducts in a year’s growth. The
position of the row of traumatic resin canals is usually constant in the same species, and
they may lie either in the spring or summer wood or in an intermediate position.

Figure 14, plate 2, brings out an interesting peculiarity in the structure of the wood
of Abies magnifica. In this species the resin canals which are normally absent in the
wood of the vegetative axis are found to occur with great constancy in the woody axis of
the cone. I have already referred to a similar phenomenon in the first memoir of this
series devoted to the anatomy of Sequoia. In S. gigantea in all the numerous cones
which have come under my notice from various sources, resin canals have always
appeared in the axis of the female cone, although they are normally quite absent from
the annual rings of the older parts of the stem and branches. I have had access to four

8 EKDWARD C. JEFFREY ON

different examples of the large cones of A. magnifica, three of which were from entirely
different sources. Rows of resin canals such as appear in figure 14, occurred in all of
them. Special care was taken to observe whether the cone or its peduncle had suffered
injury in any of these examples; but in all the cases examined both organs were quite
normal in every way. Figure 15, plate 2, shows a part of another section of the same
cone as that shown in figure 14, more highly magnified. The plane of section in this
case is lower down in the cone, and the figure as a consequence, although under a
greater magnification than the preceding one, shows the resin ducts of smaller size. The
resin ducts become fewer and of less caliber as the base of the cone is reached, until finally
in the peduncle, or even above the actual peduncle, they disappear altogether. If imstead
of progressing towards the base of the cone, one moves in the opposite direction, the
resin ducts instead of becoming fewer and of smaller size, are found to have larger lumina
and to become more numerous. ‘The larger ducts of the upper part of the cone by reason
of their greater size and close proximity tend to become fused with one another, as is
clearly shown in figure 16, plate 2. The resin secreting cells which surround the resin
ducts of the woody axis of the cone in the species under discussion are thick walled and
are quite similar to those surrounding the resin passages in the wood of Larix. Thin
walled, resiniparous cells are, however, occasionally found where a resin canal abuts
directly on a medullary ray. Although a good deal of brown tanniniferous secretion
is found in the medullary rays of the cone, little or none appears in the parenchymatous
cells surrounding the resin ducts.

In the memoir on Sequoia reference is made to the fact that resin canals occur in the
first annual ring only of vigorous shoots, in the case of trees which have already pro-
duced female cones. Interestingly enough, the same state of affairs has been found by

the writer in the case of the present species. Figure 17, plate 5, reproduces the two first

annual rings and part of the third in a cone-bearing branch of A. magnifica from the col-
lection at the Arnold arboretum of Harvard university. JI am indebted to Mr. George 8.
Shaw and Prof. Charles 8. Sargent for the opportunity of examining authentic material
of the present and other species studied in connection with the present work. The
magnification is just sufficient to show the presence of resin ducts which are entirely con-
fined to the first annual ring of the four year old shoot in question. There was no evi-
dence of any injury to the branch, so the presence of the ducts must be regarded as a
perfectly normal feature, especially in view of other similar observations shortly to be
recorded. In the stouter lateral branches of the shoot in question there was in some
instances a similar occurrence of ducts in the first annual rig; but in the more delicate
twigs these were entirely absent, thus duplicating the similar phenomenon found in
S. gigantea. Figure 18, plate 5, shows the structure of the first and second annual rings

—————

ANATOMY AND PHYLOGENY OF THE CONIFERALES. g

in a two year old branchlet from a cone-bearing branch of a different specimen of A. mag-
nifica. Resin canals are clearly present in the first annual ring and absent in the second.
No injury to the tissues could be made out in the case of this branch any more than in
the case shown in figure 17 of plate 3. In figure 19, plate 5, is shown a cross section of
the internal region of the woody ring of a vigorous branch from a flourishing specimen
of A. magnifica, between 8 and 10 meters high, growing in the Hunnewell pinetum at
Wellesley, Mass. Resin ducts are present here as well as in the other cases. In less vig-
orous branches from the lower part of the tree the resin ducts do not appear. In figure
20, plate 3, a portion of a section from the same branch as that appearing in figure 19 is
represented more highly magnified to show the nature of the resin ducts and of the cells
forming their walls. It will be seen that the ducts under consideration are quite close
to the primary wood. They are obviously surrounded by the same thick walled, resinipar-
ous cells as occur normally in all the annual rings of Larix. The magnification is suffi-
cient to show the presence of nuclei in four or five of the parietal cells of the ducts. The
nuclei present in the remaining cells have been missed by the plane of section. To sum-
marize, resin canals surrounded by thick walled, epithelial cells, such as are found about
the ducts in the annual rings of Larix, occur in the first annual ring only of vigorous
branches from vigorous trees of Abies magnifica.

In longitudinal section the pith of branches of A. magnifica presents a curious
appearance, which is shown in figure 21, plate 3. The medulla is crossed by diaphragms
consisting of sclerotic tissue, and alternating with these are segments of soft parenchyma,
which tend to break down, thus producing a fistular medulla, such as occurred in many
of the Cordaites. Sir J. William Dawson (’88) was the first one, so far as I am aware,
to call attention to this feature in the genus Abies. It is referred to here because it
appears to have a certain significance in connection with other facts to be described and
discussed at a later stage. Noticed by Dawson first in the case of A. balsamea, it seems
to be thoroughly characteristic of the genus, but occurs also in the curious Cedrus-like
species, sometimes united under the subgeneric name Keteleeria.

In figure 22, plate 5, is seen a transverse section of a root of Abies nobilis, a species
somewhat closely allied to A. magnifica. The root of the latter species differs from it
only in unimportant details. The most striking feature of this root is the presence of a
resin canal in the center of the primary wood. This peculiarity has been noted by De
Bary (’84) as being common to the genera Cedrus and Abies. To these Strasburger (’91)
adds the so called Chinese larch, Pseudolarix. The central resin canal in the case under
discussion is surrounded by several layers of delicate parenchyma cells, the innermost of
which functions as the resiniferous epithelium. There are normally no resin canals
present in the secondary wood of the root of Abies, and the examination of a consider-

10 EDWARD C. JEFFREY ON

able number of roots of different species of the genus from the Arnold arboretum of Har-
vard university and from the Hunnewell pinetum at Wellesley, Mass., has revealed no
exceptions to this rule. The cells of the phloem in the root of Abies are sometimes very
small. As an example of this feature A. magnifica may be cited. Sclerotic nests of cells
are somewhat constantly present in the so called secondary cortex of the root in Abies,
and may clearly be seen in figure 22. Resin ducts occur occasionally in the same situ-
ation, but there is never any communication between these and that which is present
in the first annual ring of the wood.

In figure 24, plate 3, 1s shown a transverse section through the woody axis of the
cone of Abies appolinis. This is the only species of the Old World in which resin canals
have been found in the wood of the reproductive axis. Two cones, which as far as could
be learned were from different sources, showed this peculiarity. In the one figured there
are two rows of resin ducts present extending completely round the cone, although some-
what better developed on one side than on the other. Figure 25, plate 4, shows a portion
of a similar section more highly magnified. The resin canals in this case do not become
very large towards the top of the cone, as in A. magnifica, but resemble those of the
former species in disappearing entirely at its base. The cells which surround the resin
canals are characterized by the presence of a small amount of tanniniferous material in
the form of minute granules, but they contain much less of it than do the cells of the
medullary rays. A. apollinis differs from A. magnifica in that the resin ducts of the
wood of the stem are confined to the axis of the cone, and do not appear in the
first annual ring even of vigorous vegetative branches. The only other species of Abies
in which resin canals have been found in the cone is A. grandis, as has been described in
the first memoir. Here, however, the occurrence of the resin passages appears to be
sporadic, for they have only been observed in a single specimen and are absent in four
others. In A. balsamea, A. fraseri, A. concolor, A. amabilis, A. nobilis, A. bracteata,
A. veitchii, and A. cephalonica no resin canals were found in the woody axis of the female
cone. A. nobilis was erroneously described in the first memoir as possessing resin ducts
in the wood of the female cone. The cone in which these structures were found turned
out to be incorrectly diagnosed, and was really A. magnifica. Keteleeria in its anatomy
obviously belongs with Abies although it has the habit of Cedrus; the wood of both vege-
tative and reproductive axes is entirely without resin canals.

The leaf trace of Abies is often double, even in the lamina, and where a geminate
fibrovascular strand cannot be distinguished in the blade of the leaf, it can usually be
made out in the part of the trace which runs in the cortex. This, for example, is the
case in Abies magnifica, where the foliar trace does not give clear evidence of being a
double strand in the free flattened portion of the leaf. By following the trace down into

ANATOMY AND PHYLOGENY OF THE CONTIFERALES. 1]

the cortex of the axis its double character can generally be clearly distinguished. Ber-
trand ('74) describes a double trace as characteristic of the genus Abies in general, and to
this statement I can subscribe. In figure 25, plate 3, is represented a foliar trace of A.
magnifica in transverse section at a height of two millimeters in the free part of the leaf.
The striking feature of this figure is the fact that the transfusion tissue, connected with
the bundle of the leaf in this as in most other Conifers, is unusually abundant, and instead
of forming a scattered reticulum about the fibrovascular strand, interspersed with living
parenchyma cells as is the rule, constitutes a dense and continuous ring of tracheidal
tissue, completely surrounding the phloem. There is a tendency for it to become inter-
rupted opposite the cambium on either side of the bundle by parenchyma cells. The con-
dition of the transfusion tissue found in this case is to be compared with that recently
described by Miss Stopes (:03) in the leaf of a Cordaite, probably the Cordaites princi-
palis of Renault. In his description of the bundles of this leaf, Renault interprets a layer
of tracheary tissue outside the characteristic centripetal wood, found so commonly in the
leaves of the more ancient Gymnosperms, as centrifugal so called phanerogamous wood,
similar to that occurring, for example, in the bundles of the petiole of the Cycads.
Miss Stopes has pointed out that this so called centrifugal wood is separated from the
centripetal xylem by a layer of irregularly disposed, thin walled tissue comparable in
every way to primary phloem. Moreover the tracheary tissue in question is attached to
the flanks of the centripetal wood, and its elements, arranged in the form of a double
sheath, correspond to the transfusion tissue of the Conifers and other Gymnosperms.

In this connection Miss Stopes calls attention to the researches of Worsdell on trans-
fusion tissue in the Gymnosperms. Worsdell (’97) reaches the conclusion that the trans-
fusion tissue, “which occurs almost universally in the leaves of Gymnospermous plants as
an auxiliary conducting system, has been phylogenetically derived from the centripetally
formed xylem of the vascular bundle.” In the example in figure 23, plate 3, the trans-
fusion tissue extends well round towards the back of the flanks of the bundle, and there
are even some elements formed in a centripetal direction. Miss Stopes draws the conclu-
sion that the sheath described by her in Cordaites principalis is not centrifugal wood, but
“a primitive stage in the development of transfusion tissue from centripetal xylem.”
The condition of affairs shown in our figure 23 seems to be strong evidence for the truth
of this hypothesis, and it supplies a transition from the condition ordinarily found in the
Conifers to that sometimes found in Cordaites as described by Miss Stopes. At a higher
level in the leaf the continuity of the transfusion tissue on the outside of the phloem is
generally lost, and we have resulting a condition in which there are merely massive
wings of transfusion tissue on the flanks of the bundle, a state of affairs which is perma-
nently present in Sequoia, Cunninghamia, etc. A good figure of this arrangement of the

iy EDWARD C. JEFFREY ON

transfusion tissue in Cunninghamia appears in De Bary’s textbook (’84). There is
another possible fate for the transfusional sheath, however. Instead of becoming inter-
rupted opposite the phloem, it may only suffer a sort of general disintegration by the
intrusion of parenchyma cells into its mass. This condition is exemplified by Pinus as

99

described by Strasburger (91, p. 153). The interesting arrangement of the transfu-
sion tissue represented in figure 25, plate 3,1is not confined to A. magnifica, but also
occurs in a less marked degree in the allied species, A. nobilis and A. bracteata. In
these species, too, the sheath tends to become discontinuous in the upper part of the

lamina of the leaf. This is especially true of A. bracteata.

PSEUDOLARIX.

Although this genus, in spite of its deciduous cone scales, is generally put in prox-
imity with Larix on account of its deciduous foliage, its anatomy as will be shown in the
sequel makes it desirable to place it with Abies.! Figure 26, plate 4, is of a nine year old
branch of Pseudolarix kaempferi. In the first six annual rings the wood closely resem-
bles that of Abies, as shown in our figure 10, plate 2. In the seventh year of growth
the stem suffered injury and as a result there is a great hypertrophy of the wood in the
region of the wound, accompanied by a corresponding reduction in the thickness in the
region of the annual rings, which is on the opposite side of the stem. In the first annual
ring formed after injury there are tangential rows of resin canals running backward from
the wound on either side. In figure 27, plate 4, some of these canals from the right hand
side of the foregomg figure are represented under a higher magnification. The canals
in this case are surrounded by cells containing more or less starch. Similar starch-bear-
ing cells occur normally as a single row on the face of the aestival wood of each annual
ring. The cortex and phloem of Pseudolarix are characterized by the presence of num-
erous mucilage cells, which become very large externally. In figures 26 and 27 they are
emptied of their mucilaginous contents, which are very soluble in the aqueous stains most
suitable for differentiating sections for photomicrography, the method of illustration used
throughout this article. There are no resin canals present in the cortex of the vegeta-
tive axes of Pseudolarix. They do not appear even in the cortical tissues of the seedling.
As has been shown above, they are equally absent in the ligneous tissues. There is this
difference between the wood and the cortex in regard to the absence of resin ducts, viz.,
that they may be produced in the wood as the result of injury, whereas they never occur
in the tissues of the cortex under this condition. Not even in the most vigorous vege-

1Since the above was written I have had the opportunity of examining male cones of Pseudolarix. The pollen has
air sacs like that of Abies and Cedrus, and in this feature presents a marked contrast to Larix.

ANATOMY AND PHYLOGENY OF THE CONIFERALES, i

W

tative shoots of the genus under consideration have I ever found resin ducts in the first
annual woody ring, such as occur in Abies magnifica as described above, and in Sequoia
gigantea described in the former memoir.

The root of Pseudolarix closely resembles that of Abies, for it has the same central
resin canal in the primary wood and numerous mucilage cells in the cortex as well. The
sclerotic nests which are characteristic of the cortical tissues of Abies are absent in the
present genus. Figure 28, plate 4, sufficiently illustrates the truth of the statements
made above. The dark cells appearing in the cortex contain tannin, while the large
empty ones are those which in life contain mucilage. In figure 29, plate 4, a part
of another root is represented. In this instance the section was treated with chrome
alum, in order to render the mucilage insoluble in the aqueous dyes employed in staining.
In one of the mucilage cells on the extreme right of the figure the contents, although
rendered insoluble by the use of chrome alum, have come out bodily, and are lying
over some of the other cells. The central resin canal, which is always present in the
root of Pseudolarix, is surrounded by parenchymatous cells with thin walls of unmodified
cellulose, and thus closely resembles that of Abies.

In figure 50, plate 4, appears a transverse section of the ligneous portion of an
injured root from a seedling from Pseudolarix. The magnification is sufficient to show
a semicircle of resin ducts running off from the margins of the wound. There is
no corresponding formation of traumatic resin canals in the cortical tissues, which
like those of the stem are normally entirely without resin ducts. I have had no
opportunity of examining the reaction of the root in the mature plant of Pseudo-
larix when injured, but there seems no reasonable doubt that the result would be
the same as in the stem of the seedling described above.

Figure 31, plate 4, shows the structure of a three year old branch of Pseudo-
larix as seen in transverse section. The material in this case came from a large
tree growing in the Hunnewell pinetum at Wellesley, Mass., which has produced cones
abundantly for a number of years. From the seed of this tree a number of seedlings
have been grown by Mr. Jackson Dawson at the Arnold arboretum of Harvard
university, so that there can be no doubt as to its vigor. There are no resin canals
present, either in the cortex or in the wood. The empty spaces seen in the former
tissue are the mucilage cells referred to above. In figure 32, plate 4, is a transverse
section through the bracteolate portion of the peduncle of a female cone of the spe-
cies under consideration. Resin canals are equally absent in both wood and cortex.
The mucilage cells of the cortex in this case are much fewer in number than they
are in the vegetative axis, and in the higher portion of the peduncle they disappear
entirely.

14 EDWARD C. JEFFREY ON

In figure 33, plate 5, is shown part of a transverse section of the axis of the female
cone in Pseudolarix. On the lower side of the figure may be seen a few small bundles
belonging to the fibrovascular ring of the cone axis. On the inner side of these is a
dense mass of sclerenchymatous tissue, which is the only part of the cone left hanging to
the tree after the very deciduous ovuliferous scales have fallen away, carrying with them
piecemeal the bundles of the axis to which their own fibrovascular supply is attached. Out-
side the ring of fibrovascular bundles, the cortical tissues of the reproductive axis are for
the most part made up of delicate parenchyma cells; but these are interspersed with a
considerable number of short stone cells, which make the cutting of suitable sections of
the axis arather difficult matter. The most interesting features of the cortex of the
cone, however, are the presence of resin canals similar to those occurring in the cortical
tissues of all the organs of most of the Abietineae, and the absence of the mucilage cells,
such as are found in the vegetative cortex of Pseudolarix. The resin canals in the pres-
ent instance are quite large in size, and do not appear at once in the base of the cone
proper. Further, they are of short length and do not intercommunicate with each other,
a fact which makes the proper infiltrating of the axis with celloidin somewhat difficult. In
figure 34, plate 5, is seen a section through the base of one of the ovuliferous scales and
its subtending bract. The ovuliferous scale is very large and only a small part of it is
shown on the upper side of the figure, while the bract is so small as to be entirely
included. The sterile bract has not yet become free from the lower surface of its ovuli-
ferous scale and does not show the presence of any resin canals. The magnification is not
sufficient to show that the fibrovascular strand of the bract is double. In the tissues of
the ovuliferous scale proper there are numerous quite typical resin canals, such as are
characteristic of the cortical tissues of the other Abietineae. These canals are continuous
with the similar canals already described in the case of the axis of the cone.

In figure 55, plate 5, isa section of the sterile bract at a pomt some distance above its
separation from the face of the ovuliferous scale. The fibrovascular bundle has shrunken
badly as the result of imperfect preservation in an insufficient amount of alcohol. The
important point to note in the present figure is the presence of a pair of resin canals, one
on each side of the gap in the tissues corresponding to the position of the fibrovascular
bundle.

Figure 36, plate 5, represents a cross section of one of the rather slender leaves of
Pseudolarix under a considerable degree of magnification. The histological elements
present are very small so it 1s not easy to make out that there are present five resin
canals on the margin of the mesophyll and over against the hypoderma, which surrounds
the leaf underneath the epidermis. Of these canals three are on the upper surface, one
of which is median and two lateral; the remaining two canals are placed laterally on the

ANATOMY AND PHYLOGENY OF THE CONIFERALES. 15

lower border of the leaf. It will thus be seen that resin ducts are present in the cortical
tissues of the leaf in Pseudolarix, as well as in the female reproductive axis, although
they are quite absent from the cortex of the stem. From the fact that there are two
resin canals only present in the leaves (7. ¢., sterile bracts) of the female reproductive
axis, it may perhaps be inferred that two is the primitive number for Pseudolarix. It
should be stated at this point that the two resin canals of the sterile bracts end blindly
below, just as do those of the ordinary vegetative leaves, in spite of the fact that the axis
to which they are conjoined, unlike the vegetative, possesses cortical resin ducts.

It will be seen from the foregoing paragraphs that resin ducts are present in the
wood of Pseudolarix under two conditions, viz., normally in the primary wood of the root
and as the result of injury in the secondary wood of both root and shoot. In the cortical
tissues they are normally present in the reproductive axis, the ovuliferous scales, the
sterile bracts, and the ordinary vegetative leaves. They do not occur in the vegetative
branches nor the root even in response to injury.

It may be noted by looking a little closely at the fibrovascular bundle of the leaf of
Pseudolarix, reproduced in our figure 36, plate 5, that the leaf trace is double, being
divided into two by a parenchymatous strand passing through the xylem. As in the
genera already described, the leaf trace in the genus under consideration is very dis-
tinctly double in the outer cortex of the axis, although it departs from the central
cylinder as a single strand as in Picea.

CEDRUS.

In the case of this genus we shall first examine Cedrus atlantica. The material
studied came from a small tree of the variety glauca growing in the Hunnewell pinetum,
at Wellesley. As is well known, those fossil woods which are without resin cells or resin
ducts, which do not possess spirally marked tracheids, and have their bordered pits
opposite and not alternating in the same tracheid, are united in the structural genus
Cedroxylon. The characters just enumerated are found in the wood of Cedrus atlantica.
Although resin canals are normally absent in the wood of this species, they occur
frequently as a result of injury under the same conditions as in the genera previously
described. Figure 37, plate 5, shows the structure of the wood on the margin of a wound
in the species being considered. Resin ducts of the usual type form a continuous row, to
the right of the area of injury. The cells surrounding the canals in this case are thick
walled and partially lignified as in the traumatic resin canals of some of the other genera
described above. They are enabled to pour out their secretion by the presence of numer-
ous pits. In figure 38, plate 5, is seen a transverse section of a wounded root of Cedrus

16 EDWARD C. JEFFREY ON

atlantica. Running in a circular fashion through the wood is a tangential row of trau-

matic resin canals. It is an interesting fact that in Cedrus the lining of the traumatic

s
resin canals of the injured root is different from that found in the stem under the same
conditions. Instead of the glandular cells in this case being thick walled and lignified
as they are in the stem, they are thin walled and of pure cellulose, and in fact strongly
resemble in their characteristics the resiniparous lining of the normal resin canals of the
wood in Pinus. Figure 59, plate 5, sufficiently illustrates the structure of the traumatic
resin ducts of the wood of the root in Cedrus atlantica. It will be seen in this highly
magnified representation of the canals, that they are lined with thin walled cells, con-
taining protoplasm and a nucleus. No resin canals are formed as a result of injury to
the cortical tissues of either the stem or the root in C. atlantica. Before leaving the
discussion of this species, reference must be made to an interesting phenomenon, which
occasionally is present in wounded roots. Ordinarily only vertical resin canals are
produced as the result of injury to the tissues of the wood in Cedrus atlantica; but in
one case which has come under my notice horizontal canals were present as well. These
started from the vertical canals and ran for a short distance radially outward, and then
ended blindly. Like the vertical resin ducts of the root they were lined with thin walled,
glandular epitheium and not with thick, lignified, resiniparous cells such as are found in
the stem. I have not found horizontal resin canals in the stem of C. atlantica; but this
is probably to be explained by the fact that my material was limited rather than to their
being non-existent, for I have observed such horizontal canals frequently in material of
Cedrus deodara, to be described below.

In figure 40, plate 5, is seen a part of a transverse section of the wood of Cedrus
deodara. Crossing the lowest annual ring in the figure and in the summer wood is to be
seen a row of resin ducts, which we may compare with those occurring as the result of
injury in C. atlantica. Although the piece of wood did not actually show the presence of
an injury in this case, there appears to be little doubt that the tangential row of resin
canals occurring here is due to a wound either above or below the region from which the
block of wood was cut. This appears extremely probable, since traumatic resin canals —
occur far above and below the actual spot of injury, as has come under my dbservation
many times in American genera of the Abietineae, and in Sequoia, in which, as has been
shown in the first memoir, similar conditions occur. It is further to be noted that the
resin ducts in C. deodara occur in the summer wood, in contrast to C. atlantica, where
they are found in the spring wood. The two species differ from each other m this respect
in the same way as do Sequoia sempervirens and S. gigantea. It is worth while to call
attention to the different position of the traumatic resin ducts of the two species of Cedrus
under discussion as it furnishes a means of distinguishing their woods. Similar differ-

ANATOMY AND PHYLOGENY OF THE CONIFERALES, 17

ences in the place of formation of the traumatic resin canals occur, as has been pointed
out above in the case of different species of Abies. Unfortunately I have not had access
to material of the wood of C. libani. The only specimens available were seedlings of the
species growing in the nursery of the Arnold arboretum, of Harvard university. None of
these showed injury, as they had been grown with great care. As a consequence they
were useless for my purpose.

Another striking feature of figure 40 is the presence of a single horizontal resin duct
passing radially outward from the row of tangential ducts. In figure 41, plate 6, is seen
another example of a horizontal resin canal from the same material. The horizontal
ducts in the two figures just described ended blindly and did not communicate externally
with other similar ducts running in the vertical plane. Often, however, when two rows
of traumatic vertical canals occur near each other in the wood, horizontal ducts extend
from the one series to the other. Such an instance is clearly shown in radial section in
our figure 42. Very often these radial ducts conjoining different tangential rows of ver-
tical ducts, are very much larger in caliber than are the vertical ducts. This feature
appears clearly in figures 42 and 45, plate 6. In figure 45, the horizontal canal extends
beyond the second series of vertical ducts only to end at a short distance to the outside of
it. It should here be stated that the horizontal ducts, so far as my observations go, always
start with a series of vertical ducts and run radially outwards, ending blindly or joining
with a more exterior series of vertical canals. They never pass inward from a vertical
series of resin ducts, which is another argument for their being entirely a traumatic phe-
nomenon. Figure 44, plate 6, is a tangential section of the wood of Cedrus deodara in a
plane outside the vertical row of resin canals shown on the left of figure 43. It will be
observed that in addition to the small linear rays, which are normally present in the genus
Cedrus, there is a number of large, so called fusiform rays. Some of these enclose a resin
canal, while others do not. The smaller fusiform rays are cross sections of resin canals
which have ended blindly, while those which are of larger size contain obvious resin ducts,
which are becoming narrower of lumen, but which have not yet been occluded.

Figure 45, plate 6, shows part of the transverse section of the axis of a cone of
Cedrus atlantica. In the row of bundles which occurs on the upper side of the figure
there are no resin canals in the woody tissues. The cavities, appearing on the outer sides
of the wood bundles, indicate the position of the collapsed phloem, the section being made
in this instance from dry material. Resin canals occur in the cortex only of the female
cone of C. atlantica. Cones of C. libani and C. deodara were examined with similar
results. In no case have I found resin ducts in the woody tissues of the cone in the genus
Cedrus."

1 Radais (°94) has, however, described resin canals in the first year’s growth of the cone of Cedrus deodara.

18 EDWARD C. JEFFREY ON

In figure 46, plate 6, is a cross section of the root of Cedrus atlantica. The
center of the primary wood is occupied by a resin canal of considerable size, which
is surrounded, as in Pseudolarix and Abies, by thin walled cells of the secretory paren-
chyma. The walls of the gland cells in Cedrus are of pure cellulose. The presence
of a median resin canal in the root of this genus has already been described by
De Bary (84). The secondary wood of the root in the various species of Cedrus con-
tains no resin ducts except in the case of injury. There are no sclerifications either
in the bast or in the cortex, and mucilage cells are absent as well. In the smaller
roots which have been mainly examined there were no cortical resin canals, although
these are abundant in the stem.

The leaf trace of Cedrus is described by Bertrand (’74) as showing some evi-
dence of duplication by the presence of a very distinct median medullary ray. In
the cortical course of the leaf trace it is obviously double, as in the case of the other
Abietineous genera described above.

This condition is shown in figure 47, plate 6, which represents the foliar trace of
Cedrus atlantica as it is passing through the cortical tissues on its way to the leaf.
A similar state of affairs has been made out in the leaves of seedlings of Cedrus libani.
It will thus be seen that in its outgoing leaf trace Cedrus is no exception to the
condition of affairs described above for the other Abietineae. Although the foliar
trace in this genus is geminate in the middle and outer cortex, it nevertheless
leaves the central cylinder of the axis as a single strand.

TSUGA.

In figure 48, plate 6, is shown a transverse section of an injured stem of Zsuga
canadensis. There are four normal rings of wood present and part of another. The
injury took place obviously in the early summer of the fifth year of growth of the
branch represented in our figure. From the mode of formation of the traumatic
wood it is clear that the injury must have been very severe, for the first ring
of woody tissue formed subsequently to this event extends only half way around
the stem. Although the magnification is not great, it is obvious that resin canals
are present in the usual tangential row, which is characteristic of the traumatic
canals in the Abietineae in general. The resin ducts are situated in the early
formed part of the traumatic wood. The usual position of the resin ducts in
the injured wood of 7. canadensis is in the later formed portion of the spring
wood. This is clearly seen in the case of large wounds in thick stems, where a
number of years is necessary for the closing over of the exposed surface by

te

ANATOMY AND PHYLOGENY OF THE CONIFERALES. 19

traumatic wood. Under these circumstances tangential rows of resin canals are
formed in the spring wood of successive but generally not consecutive annual rings.
Although a good deal of material has been examined from various sources and of
wide geographical distribution, no examples of horizontal resin canals have been found
such as are present in Cedrus, described above. The resiniparous cells lining the
interior of the traumatic resin ducts in Tsuga are thick walled and much pitted.
They frequently contain dark brown tanniniferous globules or masses, similar to
those found in the resin cells of the normal wood of Tsuga. Traumatic resin canals
of a similar nature have also been found in 7'’swga mertensiana. In this species they
are entirely confined to the aestival wood. It is thus possible to distinguish the two
species by the reaction of their wood to injury.

In figure 49, plate 7, appears a transverse section of a yearling stem of Tsuga
canadensis. Although the branch is a vigorous one there is no evidence of the presence
of resin ducts. Even in the most flourishing branches of 7. canadensis, bearing female
cones, there are no resin canals present in the first year’s growth, such as are
found in similar shoots of Abies magnifica. Another feature of Tsuga shown in the
figure is the entire absence of cortical resin canals. This condition is likewise present
in T. mertensiana, which resembles 7. canadensis in the structure of its woody tissues.
The cortex of the two species mentioned is also free from the mucilage cells which
are such a striking feature of this tissue in Abies and Pseudolarix.

In figure 50, plate 7, is shown part of a transverse section of the cone of Tsuga
mertensiana. Although there are still no resin ducts in the wood of the large
cone of this species, they are obviously present in the cortex. On the right of
the figure two very large ducts are about to pass off into the fructiferous scale. Figure
51, plate 7, shows a similar view of a transverse section of the cone axis of 7.
canadensis. In this case, too, cortical resin canals are present although there are no
ligneous ones. On the right and left of the figure, ducts are to be seen passing
out into the two fructiferous scales which are included in the plane of section.
It will be seen that Tsuga resembles Pseudolarix in the distribution of cortical
resin canals in the vegetative and female reproductive axis, for as in the latter
genus, they are present in the reproductive axis and in the fructiferous scales,
but are entirely absent from the cortex of the vegetative stem. As is well known,
there is a single median resin canal present in the mesophyll of the vegetative leaf
of Tsuga.

In figure 52, plate 7, is represented a transverse section of the root of Tsuga
mertensiana. The magnification is just sufficient to show the presence of a median
resin duct in the primary wood. There are no resin canals present either in the

20 EDWARD C. JEFFREY ON

secondary wood or in the secondary cortex which appear in the figure. Figure 52
shows the structure of a three year old root of Tsuga canadensis. There is the same
median resin duct as in the preceding figure, but no resin canals appear in the second-
ary wood, the phloem, or the cortex. In figure 54, plate 7, is shown a section of a
wounded root of 7. canadensis. There is a row of resin canals present in the trau-
matic wood formed subsequently to the receiving of the injury. The cells surrounding
these canals are of the thick walled type, and they often contain considerable tanninifer-
ous substance. Figure 55, plate 7, reproduces the structure of the central portion of
the primary wood and part of the secondary wood which les about it. The parenchyma
separating the primary from the secondary wood contains a large amount of tannin, but
the cells surrounding the resin canal in the center of the primary wood are obviously
free from this substance. The resiniparous lining of the median resin canal of Tsuga
is thicker walled than in Pseudolarix, Abies, and Cedrus, but its cells are nevertheless
quite typical resin-secreting elements. In figure 56, plate 7, is shown the passage of
a leaf trace into the leaf in Z'’suga mertensiana. The trace is plainly double in this
instance, as in the other Abietineae described in this investigation. Bertrand (’74) cites
Tsuga as a genus possessing a single leaf trace. If the exit of the trace from the central
cylinder be studied in this genus, it becomes evident that the fibrovascular strand is
clearly duplicate even before it leaves the wood of the axis, and persists as an obviously
geminate strand up into the base of the leaf proper. T’suga canadensis was examined
with the same result, although here the size of the strands of the double trace is some-
what smaller. With the description of the leaf trace in Tsuga, the consideration of the
recognized genera of the Abietineae, with the exception of Pinus, is completed. In the
following paragraphs the bearing of the observed facts on the phylogeny of the order

will be discussed.

CONCLUSIONS.

Having examined from a comparative standpoint the general features of the anat-
omy of the Abietineae, we are now in a position to discuss the significance of the facts
described, in connection with the affinities and phylogeny of the order. The discussion
may conveniently be divided under three heads: (1) The bearing of anatomy in the
broadest sense on the interrelationships of the various genera of the Abietineae; (2)
The relationship of the Abietineae to the other groups of the Coniferales; (3) The affinity
of the Abietineae with other groups of Gymnosperms, living and extinct. These will be
considered in the order in which they have been mentioned.

The interrelationships of the Abietineae—In the foregoing paragraphs it has been

ANATOMY AND PHYLOGENY OF THE CONIFERALES. 2]

noted that certain genera of the Abietineae were characterized by the abundant occur-
rence of resin canals in the secondary wood of their vegetative and reproductive axes and
in the secondary wood of the root as well. In these genera resin canals are also abun-
dant in the cortical tissues of all the organs. They are further characterized by their per-
sistent cone scales.

In contrast to these are a number of genera in which the resin canals, so abundant in
_the forms just described, are almost entirely absent in the secondary woody tissues of the
vegetative and reproductive axes. Where they do occur in the secondary wood of the
stem organs, they are confined to the female reproductive axis, and to the first annual ring
of vigorous branches of sexually mature trees. These genera are further particularized
by the fact that in the woody tissues of the root, resin canals are confined to the primary
wood, but are never absent from this position. In the cortical tissues of some of the
genera included in this group, resin canals are abundantly present, but in others they
occur only in the female reproductive axis and its appendages, and in the mesophyll of the
leaves. Although resin ducts are normally absent in the secondary wovod of the vege-
tative axial organs of the genera under discussion, their formation may be brought about
in dense tangential rows by the attacks of fungi, or by mechanical injury. Finally the
genera included here are with one exception characterized by deciduous cone scales.

The genera to be included in the first group are Pinus, Pseudotsuga, Larix, and
Picea. The second group contains the genera Abies, Pseudolarix, Cedrus, and Tsuga.
It seems advisable to include the first four genera in a special subfamily, the Pineae, and
the last four in another special subfamily, the Abieteae. These two series are so distinct
in both their reproductive and anatomical characters, that it will conduce to a clear under-
standing of their phylogenetic relationships to keep them separate under the designations
which have been suggested above.

Having separated the genera of the Abietineae into two distinct groups we may now
go on to inquire how these two groups are related to each other and which of them is the
more primitive. Further, after having as far as possible from the data here made avail-
able, settled their general relationship, we may advantageously pass on to the consider-
ation of the affinities of the several genera which they include.

As a preliminary to the discussion of the relative antiquity of the Pineae and
Abieteae as here defined, it will be convenient to recall the anatomical differences
between the two subfamilies. The Pineae are characterized by the presence of resin
ducts in the woody and cortical tissues of their stems and roots. There are peripheral
resin ducts in the primary wood of the root and similar ducts are in some cases found in
the primary wood of the stem. The integument of the ovule is relatively thin and the
scales of the cone are not deciduous. In the Abieteae on the other hand, resin canals are

22 EDWARD C. JEFFREY ON

entirely absent in the woody tissues of the stem. The only exception to this rule is
their occasional occurrence in the wood of the reproductive axis (species of Abies) and
in the first annual ring of vigorous branches in mature trees (species of Abies). Resin
ducts occur normally and without exception in this group in the center of the primary
wood of the root. Resin canals are also present in the cortex but in the genera Tsuga
and Pseudolarix the cortical canals are confined to the reproductive organs and to the
leaves. In the Abieteae the integument of the ovule is relatively thick, and the cone
scales are usually deciduous.

As a further preliminary to the discussion of the phylogeny of the Abietineae
attention must be called to certain principles or canons of comparative anatomy, derived
from the study of living and fossil plants, which are of such recent origin that they
have scarcely as yet been definitely formulated. One of the most important of these
canons is, that ancestral structures are apt to linger on in the reproductive axes after
they have disappeared from the vegetative organs. Solms-Laubach (’90) first called
attention to this principle in the matter of the arrangement of the bundles in the
cones of the Cycads, which is much simpler in the reproductive axis than in the vege-
tative stem, and according to his view is archaic. The most important contribution
to this generalization, however, has been made by Scott ('97). He poimted out that
in certain Cycads there are present, in the axis of the cone, mesarch bundles compar-
able to those found in the vegetative axes of certain of the fossil and ancestral
Cycadofilices. He draws the inference that the mesarch type of bundle is primitive
for the Cycads and has lingered on in certain instances as an anatomical feature of
the male and female cones. The present writer (’99) has directed attention to the fact
that the course of the fibrovascular bundles in the cones of the more modern Cala-
mites is the same as that found in both vegetative and reproductive axes of the more
ancient Calamites included in the genus Archaeocalamites; for in the latter there is
no alternation of the course of the bundles at the nodes, a feature retained in the
cones of the later Calamites and to some extent even in the modern Equisetum. ~ In
the younger stages of the development of the fibrovascular bundles of the cone in
Equisetum there is present a cambium, another point of resemblance to the Cala-
mites, ancient and less ancient, which were characterized by the presence of very
marked secondary woody growth. In this connection it may be pointed out that
Cushman (:03) has noted the resemblance of the leaves occurring on the reproductive
axis of certain higher plants to those first appearing in the seedlings. Jackson ('99)
had previously shown that the nepionic or earlier leaves of phaenogamous seedlings
in many cases present remarkable resemblance to the leaves of allied fossil species.
These illustrations will serve to indicate that there is a strong tendency towards the
retention of ancestral characteristics in the reproductive axes of vascular plants.

ANATOMY AND PHYLOGENY OF THE CONIFERALES. 23

Persistent as ancestral characters are in the reproductive axis, they are much
more markedly so in the foliar organs. In the case of the Cycads, the fibrovascular
bundles of the leaves are of the old quasi-cryptogamic mesarch type found in the
medullary crown of the vegetative stem of Lyginodendron and others of the transitional
Gymnosperms. We thus find living on in the leaf, a type of bundle which in the
Cycads has long disappeared in the vegetative stem and occurs only in rare instances
even in the more conservative reproductive axis (Scott, 97). A similar feature is
presented by the leaf bundles of Lyginodendreae (Williamson and Scott, ’95) and the
Osmundaceae (Faull, :01; Jeffrey, :02), for in these groups the concentric Filicinean
type of fibrovascular strand, which has almost disappeared in the fibrovascular zone
of the stem, persists, often in full vigor, in the leaves. It has further been shown by
Jackson (99) that the outward form of the earlier or nepionic leaves of many Angio-
spermous and Coniferous seedlings presents striking resemblances to the adult leaves
of extinct fossil species. Numerous other examples and illustrations of the same
principle might be cited, but the foregoing have been chosen because they are
particularly convincing on account of their confirmation by fossil data.

Lastly, in this connection we come to the ancestral features presented by the
seedling. It is a recognized principle in the embryology of animals that ancestral
features are apt to persist in the earlier stages even where they have quite dis-
appeared in the adult. As a striking illustration of this principle may be cited
the fact that gill slits are found in the embryos of all the great groups of the Ver-
tebrata, although gills have long disappeared in the adult of most of the classes
of the phylum. In the case of plants, as has been pointed out by Scott ('96),
this feature is much less marked, but striking examples of recapitulation in the
case of seedlings have been described by Strasburger (’72), Goebel (’89, date of part),
Jackson, and others. The Cupressineae present particularly good illustrations of this
principle, since in them the older stem has undergone very marked xerophytic modi-
fications in the way of flattened, appressed leaves, etc. These extreme xerophytic fea-
tures are absent as a rule in the seedling, where a somewhat Taxodineous type of
foliage is often found. The more ancestral type of foliage may further be discov-
ered in the relatively more vigorous main axes, where there is often present a
larger and less appressed type of leaf. In this connection may be mentioned the
reversion to ancestral conditions of internal structure, which sometimes occurs in the
first annual ring of vigorous branches. The writer has pointed out in the first
memoir of this series (Jeffrey, :03) that in the more vigorous ramifications of the
cone-bearing shoots of Sequoia gigantea there occur, in the first annual ring only,
resin canals such as are found as an unfailing structural feature of the wood in

24 EDWARD C. JEFFREY ON

the female cone. This feature is the more remarkable since resin canals are absent
ordinarily in the woody tissues of the Sequoias and are replaced by resin cells of
the usual Cupressineous type. This characteristic of the first annual ring of vigorous
branches of mature trees of Sequoia gigantea must apparently be interpreted as ances-
tral, since it is paralleled by the presence of a similar feature in the reproductive axis.’
The condition of reduction from a more luxuriant ancestry also explains why the
resin canals in question are not found in the seedling in the first annual ring.
Apparently there must be a considerable accumulation of nutritive capital before
the species under discussion is able to reproduce the ancestral conditions even in
the first annual ring. It may be set down then as a rule, which in the future will
doubtless receive numerous exemplifications, that in forms which have clearly suffered
vegetative reduction the reappearance of ancestral conditions may be delayed to the
first annual ring of the more vigorous shoots of the adult.

A further general principle of comparative anatomy, which is beginning to be estab-
lished, is that ancestral conditions may be recalled by appropriate experimental conditions
Where there has been reduction from a more luxuriant ancestral type, experimental con-
ditions, which bring about a greater determination of nutrition to the parts im question
are apt to cause a reversion to the ancestral structure. It has been shown by Jackson
(99) that this is often true of the external form. In the case of Sequoia, as I have
pointed out in the first memoir of this series (Jeffrey, :03), injuries to the phloem and
the cambium, which bring about a damming up of the products of assimilation in the
region of the injury, cause a hypertrophy of the woody tissues. In this hypertrophied
wood more or less typical resin canals occur, similar to those found in the first annual ring
of certain branches and invariably in the woody tissues of the female cone of S. gigantea.
It is apparent that experimental morphology is likely in the near future, to receive from
botanists the same attention that it has already begun to enjoy in the hands of the
zoologists.

The phylogeny of the Abietineae.—The enunciation of general principles briefly
attempted in the foregoing paragraphs will serve a useful purpose in the discussion of the
phylogenetic relationships of the Pineae and Abieteae. In the latter group, it has been
noted in the descriptive part of the present memoir that resin canals are generally absent
from the woody tissues of the root and shoot. Exceptions to this statement occur in
the following cases. In the genus Abies, resin ducts have been found in the wood of
the cones of A. grandis, A. magnifica, and A. apollinis. In A. magnifica (as in
Sequoia gigantea) these are also in the first annual ring only of vigorous vegetative
branches of mature trees. Resin ducts also are found invariably in the primary wood

1 Radais (’94) has also noticed this feature of the cone in S. gigantea.

ANATOMY AND PHYLOGENY OF THE CONIFERALES. 25

of the first annual ring of the root. The resin canals present under the conditions
described above are perfectly normal, and are surrounded by a continuous jacket of
resiniparous epithelium. A further mode of occurrence of resin canals in Abies is
that seen in the case of injury to the bast and cambium, either from mechanical causes,
or by the attack of parasitic fungi. The resin canals formed under these conditions are
generally confined to the wood, and occur as tangential rows of so called “imperfect
canals,” which are in full communication with each other tangentially, and thus serve
to pour resin over the surface of the wound. The resiniparous epithelium in this
ease is ordinarily thick walled and as a consequence conspicuously pitted. The
thickening of the wall seems to be merely a mechanical device, as all parenchyma of
the wood is more or less characterized by this feature in the vicinity of a wound.
Further, the resiniparous cells surrounding the traumatic resin ducts under consideration
are apt to partake of the general features of the parenchyma normally occurring in the
wood and the medullary rays. They may often as a consequence contain the globules
of tanniniferous material, which are so constant a characteristic of the so called resin cells
in Abies and other genera of the Abieteae as limited above. In the remaining genera,
Cedrus, Pseudolarix, and Tsuga, the normal occurrence of ligneous resin canals is confined
to the first annual ring of the root. Traumatic resin ducts are found in the last men-
tioned genera as a feature of the wood in the case of injury just asin Abies. It is of
importance to note, that in Cedrus, traumatic resin canals may occur both in the hori-
zontal and in the vertical planes, with correlation between the two systems. Only
vertical ducts have been made out in the other genera of the Abieteae.

Before going on to discuss the significance of the structural features of the wood
summarized in the foregoing paragraph we may turn our attention to the distribution of
cortical resin canals in the Abieteae. In Abies and Cedrus, resin canals occur abundantly
in the cortex of both vegetative and reproductive axes and in the mesophyll of the leaf as
well. In Pseudolarix and Tsuga on the other hand, the resin canals of the cortex are
strictly confined to the female reproductive axis together with its appendages and to the
mesophyll of the leaves. Cortical resin ducts cannot be produced as the result of injury.

The distribution of the resin canals in the wood and cortex of the Abieteae is such
that if the validity of the general principles enunciated above be admitted, we must
regard them as having come from ancestors having resin canals throughout their wood
and cortex, as is the case with the genera included in this memoir in the subfamily
Pineae. The resin canals in the Abieteae have almost entirely disappeared in the woody
tissues, only retaining their place in an isolated fashion in such conservative regions as the
wood of the female cone and the first annual ring of the root and shoot. They further
show their vestigial character in the phenomena of injury. In the cortical tissues there

26 EDWARD C. JEFFREY ON

is a close parallelism of distribution, for in Tsuga and Pseudolarix, in which cortical resin
passages are largely obsolete, they persist in the ancestral situations presented by the
cortex of the female cone and its appendages, and the mesophyll of the leaf. If the
general principles of comparative anatomy enunciated at the beginning of this discussion
are correctly formulated the Abieteae have come from a stock characterized by the
presence in both fundamental and fibrovascular tissues of resin ducts which form a con-
tinuous system of horizontal and vertical canals, such as are found in the living Pineae.

If we attempt to picture to ourselves the probable course of evolution which has led
to the more or less complete loss of the resin canals in the Abieteae (as defined above),
it would appear to be as follows. The ancestral forms provided with a comprehensive
and freely anastomosing system of resin canals in both cortical and ligneous tissues, were
thus safeguarded against infection in case of injury, but at great cost both in the large
supply of resinous secretion necessary to supply the needs of this extensive system, and in
the large quantity expended in sterilizing a wound. On account of the reduced foliage of
even the Abietineous Conifers, this was a very serious drain on the assimilatory apparatus.
Gradually the more economical tendency arose of forming resin passages in the case of
need only. In Pinus this tendency is scarcely observable, while in the other three genera
of the Pineae it has become quite marked. In the Abieteae it has passed beyond the
stage even of a marked tendency and has become the rule, so that in this subfamily the
original extensive and anastomosing system of resin canals has become reduced to isolated
and uncommunicating ducts, quite useless as a rule from the protective standpoint and
persisting as ancestral relics in the more conservative organs and parts of the plant.
The place of this system is taken by the much less costly expedient of resin cells and
by traumatic resin ducts which are formed in case of need only.

It should be pointed out here, although this matter is reserved for fuller discussion in
connection with the study of the genus Pinus in a subsequent memoir, that this con-
clusion is supported by other anatomical considerations. A characteristic feature of
structure in the medullary rays of many of the Abietineae is the presence of marginal
tracheidal cells, generally interpreted as favoring the radial movement of water in the
wood and rendered necessary by the fact that the pits of the tracheids of the wood are
mainly on the radial walls, thus providing for a tangential rather than a radial movement
of water. These marginal tracheidal cells are well marked in the genera Pinus, Picea,
Larix, and Pseudotsuga, but as has been recently noted by Strasburger are much less
prominent in Tsuga and Cedrus, while as was long ago pointed out by De Bary, in Abies
they are lacking except in a few species. My own observations in the case of Pseudo-
larix show them to be absent here, although I have not had access to sufficiently robust
stems to be able to state positively that they are never present in this monotypic genus.

ANATOMY AND PHYLOGENY OF THE CONIFERALES. 27

Penhallow (:03) assumes that the occurrence of these ray tracheids marks a high stage
of evolution in the genera which are characterized by their presence. It should be

remembered in this connection, as Strasburger (’72) has emphasized, that the Conif-
erales are a group which has undergone regressive metamorphosis and as a consequence

structural features of high complexity are more apt to be characteristic of the ancestral
forms than of those of more modern origin. Further, Penhallow (:03) himself has called
attention to the fact of the sporadic occurrence of ray tracheids in the genera Juniperus
(J. communis, var. alpina) and Cupressus (C. thyoides and nootkatensis) among the
Cupressineae. Their presence here is to be interpreted rather as vestigial than rudi-
mentary especially in view of the peculiarities of the seedlings in this group, which seem
to show clearly that it is a reduced order of the Coniferales. Furthermore, the occur-
rence of ray tracheids is in itself by no means an indication that the Abietineae are a
modern order, for the same feature was present in that very ancient family, the Lepido-
dendraceae, which, as has been pointed out by Corda (Williamson, ’81, p. 294) and Scott
(:00), was characterized by the presence of scalariform tracheids in the medullary rays.

We may now pass on to consider the closer affinities of the several genera of the
Abietineae. As has been pointed out above they fall naturally into two series as a result
of the consideration both of their anatomy and of their reproductive organs. The genus
Pinus appears to stand by itself among the Pineae, for its fascicled leaves and altogether
unique cones, as well as its great geological age, separate it clearly from the other three
genera. Picea and Pseudotsuga seem to be rather close to one another, but Larix cannot
be regarded as far removed. The genus Abies includes perhaps the greatest number of
primitive features among the Abieteae. Closely allied to it is Pseudolarix, not only
in its anatomical structure and its rapidly deciduous cone scales but also in its thickly
integumented seeds. The deciduous foliage of Pseudolarix is by no means sufficient to
unite it with Larix, as is generally assumed by the taxonomist, for Cedrus, which has
always been regarded as possessing close affinities with Abies, is also occasionally decidu-
ous, as has been pointed out by Carriére and Kent (:00). Abies and Pseudolarix closely
resemble each other in the possession of very characteristic mucilage cells, which are
quite absent in Cedrus and Tsuga. Cedrus is generally conceded to be allied to Abies,
to which, as has been shown above, it presents a very close resemblance in the details
of its anatomy, the structure of its seeds, and the deciduousness of its cone scales.
It differs from Abies and Pseudolarix in the absence of mucilage cells. Of the four
genera of the Abieteae, Tsuga shows the least affinity with the others. Its cone scales
are not deciduous and its seeds are not relatively as thick walled as those of the three
first mentioned genera. Further, in the single resin canal of the leaf, it resembles many
of the Cupressineae.

28 EDWARD C. JEFFREY ON

Relationship of the Abietineae to other Coniferous orders.— It is perhaps the
prevailing tendency at the present time to regard the Abietineae as the most modern
of the Coniferous orders. This view does not appear to be supported, however, by
the recent results of palaeontological research, or by those of anatomy and general
morphology. It has been recently shown by Zeiller (:04) that Pinus, which is often
regarded as a very modern representative of the Conifers, was already present in the
Jurassic in the two sections representative of the hard (Pinaster) and the soft
(Strobus) pines. By this discovery the genus appears as a very ancient one, for if
its two distinctive sections were already well marked as early as the Jurassic period,
its primitive history must lie far in the geological past, extending perhaps to the Permian,
for Penhallow (:00) has described a Pityoxylon of this age.

In discussing the relationship of the Abietineae to the other Coniferous orders, it
will be convenient at the present time to confine our attention to the Cupressineae, in
the larger sense, for these have already been touched on in the first memoir. As has
been pointed out by Zeiller (:00), the Cupressineae are relatively modern in their
appearance in the geological strata, and it may accordingly be suggested as probable
that they are less primitive than the Abietineae. Anatomically the wood of the
Cupressineae is characterized by the absence of resin canals, except in certain sporadic
instances, and the place of these is taken by resin cells. By a course of reasoning similar
to that adopted in the present memoir, the writer has reached the conclusion that the
resin ducts described by him in the first annual ring of branches and in the wood of the
cone-axis and scales of Sequoia are an ancestral feature. This conclusion was fortified
by the occurrence of resin canals in this genus as the result of injury, and presumably
as a reversion to the ancestral condition of the wood. The inference was drawn in the
first memoir that the presence of what were apparently vestigial resin canals in Sequoia
pointed to the derivation of that genus from a stock characterized by the presence of
ligneous resin canals. The attempt has been made in the present memoir to show that
this condition of the wood is not only characteristic of the Abietineae, but is also a very
old one for this order. On the basis of wood structure alone it would appear that the
Abietineae are older than the Cupressineae. It is obviously undesirable to interpret
affinities from the structure of the wood alone, useful ‘as this has proved in tracing
affinities in the case of fossils, where no other evidence has been available.

The structure of the female cone is universally admitted to be of great importance
in arriving at the relationship of the various orders of the Coniferales. There
are several morphological interpretations of the female cone, the best established of
which explains it as a main axis bearing a number of flattened and reduced ovuliferous
shoots, each borne in the most primitive condition, in the axil of a bract. In the

ANATOMY AND PHYLOGENY OF THE CONIFERALES. 29

less primitive types of female cone the bract and flattened axis become more or less
completely fused with each other, so that in many cases it is only possible to dis-
tinguish them by their still separate fibrovascular supply. This view has received
strong confirmation. As is well known, in the genus Sciadopitys the apparent vege-
tative leaves are really two fused leaves belonging to an abortive branch and are
axillary to a scale-like bract. This interpretation is generally accepted in the
case of the genus under discussion, and as has recently been pointed out with great
aptness by Coulter (:01) is a clear case of parallel reduction in a vegetative shoot
corresponding to that hypothetically found in the case of the ovule-bearing scales of the
Coniferous cone. This parallel drawn by Professor Coulter receives the strongest con-
firmation from proliferous female cones of Sciadopitys, in which the ovuliferous scales
are actually replaced by green double needles, such as are found on the vegetative
axes of the same species (for a good original figure of one of these interesting pro-
liferous cones, see Veitch’s Manual of the Coniferae, 1900, p. 54). A shoot organ has
appeared so often and in so many different genera of the Coniferales in place of the
ovuliferous scale, that we may safely regard the two structures as homologous. If it
be admitted that the ovuliferous scale is the representative of a reduced shoot
occurring in the axil of a bract-like leaf, it follows, other things being equal, that
the group which most nearly realizes the ancestral condition will be the most primitive,
and those in which the bract and ovuliferous scale are wholly or partially fused
must be more modern and specialized. On this basis we are compelled to regard
the Abietineae, in which the bract is without exception quite free from the ovuliferous
scale, as more primitive than the Cupressineae (in the large sense) in which the bract
and ovuliferous scale are always intimately fused together,

But it is not in the structure of the female cone alone that the Abietineae appear
to show themselves more primitive than the Cupressineae. The prothallium found
in the ovule of this order is throughout characterized by the thick megaspore coat
which surrounds it. This is a feature which the Abietineae possess in common
with the Cycads and the fossil Gymnosperms. The thick coat surrounding the mega-
spore in the Cycads has been described by Warming and more recently by Lang.
The figures of Williamson, Scott, and Oliver sufficiently establish the presence of a simi-
lar thick megaspore coat in several fossil Gymnosperms. In the Cupressineae, the Tax-
ineae, the Gnetales, etc., the coat surrounding the megaspore is thin or obsolete.
The Abietineae thus present in the structure of their megaspore coat a very significant
feature of resemblance to the older Gymnosperms.

The structure of the male gametophyte in the Abietineae is also indicative of
their being more primitive and less reduced than the Cupressineae. In the first

30 EDWARD C. JEFFREY ON

order two or three prothallial cells are formed within the germinating pollen grain
previous to anthesis, as has been shown by Strasburger and others in the case of Larix,
Pinus, Tsuga, etc. In the Cupressineae, using the term in the broader sense to include
such forms as Sequoia and Taxodium as well as Thuja, Cupressus, and Juniperus,
this is not the case as only the stalk and antheridial cells and the tube nucleus make
their appearance in the germinating microspore. It may be urged that the winged
pollen, found in Pinus, Abies, Cedrus, Picea, etc., is to be regarded as a mark of high
specialization and as a consequence an indication of a modern origin for the group under
discussion. This is by no means necessarily the case, for examples of winged micro-
spores are known to have existed among the Lepidodendrids (Scott, 98), and there
are no cases of winged pollen grains among anemophilous Angiosperms, which may
be taken as the most modern type of wind-fertilized vascular plants. It seems probable
that the air sacs are for the purpose of reducing the specific gravity of the ancestrally
heavy and multicellular type of pollen grain found in the Abietineae.

Perhaps the most important indication of the primitive nature of the Abietineae
is supplied by the study of their leaf trace. It is now a well established general prin-
ciple that ancestral features are apt to persist for a long time in the vegetative leaves
of the vascular plants. This principle is, for example, very well illustrated by the
case of the Cycads, referred to in an earlier paragraph. Attention has been called
in the descriptive part of this memoir to the fact that no matter what may be the
character of the foliar strand in the lamina of the leaf in the Abietineae, it is always
a double strand in the cortex. A double leaf trace was characteristic of Lyginoden-
dron, Poroxylon, Calamopitys, the Cordaites, and others of the fossil and transitional
Gymnosperms. It is also exemplified in the foliar trace of Ginkgo and in the double
strand, which often passes off to the sporophylls in the Cycads. Scott (:00) has called
attention to this feature of the older Gymnosperms and suggests that it may have
a taxonomic value. In view of the facts described in this memoir for the Abietineae,
the double trace appears to have a peculiar significance, especially when it is recalled that
the Cupressineae have but a single foliar strand (Bertrand, 74). The topography of the
double foliar trace in the Abietineae is such that we may regard the single leaf trace,
which is found in what we believe to be the less ancient orders of the Coniferales, 7. ¢.,
the Cupressineae, Taxineae, etc., as originating by the progressive reduction in size of
the leaves. The process is centripetal, that is, it progresses from the apex of the leat
downwards into the cortex of the stem, so that starting with a single bundle in the foliar
apex we arrive sooner or later at the primitive double bundle. The separation may take
place high up in the lamina of the leaf, as has been pointed out by Strasburger (’91)
in the case of Pinus. Further, in very small leaves of the Abietineae, as has been noticed

ANATOMY AND PHYLOGENY OF THE CONIFERALES. 31

by Thomas (’66), there may be but a single strand throughout. This author is of the opin-
ion that the single bundle is more primitive, and cites Ginkgo as an example in which the
increased development of the leaf leads to the appearance of a double foliar trace. With
our present knowledge of the more ancient and especially of the fossil Gymnosperms, this
interpretation cannot be accepted, for we have now every reason to believe that the
ancestors of the Coniferales had much larger foliar organs, much more richly supplied
with strands of fibrovascular tissue, than their descendants. The progression is con-
sequently in the opposite direction, 7. ¢., the more modern Gymnosperms have suffered
reduction both in leaf surface and in the fibrovascular tissue constituting the leaf traces.
The simplification of the leaf traces into a single strand has apparently been from above
downwards. In further support of this view may be cited the observation made by
Scott (97) in the case of the reduced strands occurring in the peduncular portion of the
cone in certain species of Cycads. Here paired strands coming off from the central
cylinder of the cone, run a greater or less distance upward into the cortex and then
become fused together. Dr. Scott considers it probable that these double strands repre-
sent abortive leaves. It is interesting in connection with the present argument that
there should be the same fusion of originally separate leaf traces at the apex in the case
of reduction.

Other indications of primitive affinity, less important it is true, because less wide
spread and constant, are the occurrence of a chambered pith in Abies and in certain
species of Pinus, and in the case of the leaf trace of the former genus, the complete
enclosure of the phloem of the laminar bundle of the leaf in certain instances, by a
cordon of transfusion tissue. The first of these features was very prevalent in the
Cordaites although not universally present (Scott, :02), while the second has also been
recently described in a Cordaitean leaf (Stopes, :03).

If the arguments advanced in the foregoing paragraphs are sound, we may assume,
because they present many features of structure which unite them both with the
older Gymnosperms on the one hand, and by an obvious process of reduction, with the
Cupressineae, etc., on the other hand, that the Abietineae are a very ancient family or
order of the Coniferales. This conclusion is favored by what is known at the present time
of their geological occurrence, and will probably receive fuller confirmation from this
source when the geological record is more complete than it is at the present time.

The affinity of the Abietineae with other cohorts of Gymnosperms.— The discussion
of the relationship of the Abietineae with other great groups of Gymnosperms has been
largely anticipated in the paragraphs of the last section. Both Coulter and Chamberlain
(:01) and Scott (:00) among recent writers on the subject of the phylogeny of the
Gymnosperms have derived the Coniferales as a whole from a Cordaitean stock or plexus.

Bye EDWARD C. JEFFREY ON

The results of the present investigation seem only to add confirmation to this view.
Although the reproductive apparatus of the Cordaites, as far as we are at present
acquainted with it, was somewhat different in its organization from that of the Conifer-
ales, still as Dr. Scott (:00) has pointed out in the case of the female axis, it is not diffi-
cult to trace homologies with that of the latter group, if we regard the ovuliferous
apparatus in the Conifers as consisting of a reduced and modified axillary shoot. The
anatomical study of the vegetative organs of the Coniferales, as represented in the pres-
ent instance by the Abietineae, serves strongly to confirm the relationship of the two
groups. It seems quite impossible to refer the Coniferales any longer to a Lycopodineous
ancestry, although their superficial features seem somewhat to favor such a derivation.
It has been pointed out by the writer (:02) that the mode of exit of the leaf trace from
the central cylinder, even in those Conifers which have their foliage most reduced, is
distinctly Filicinean, and of a type which is never found in the Lycopods or their allies.
The present research increases the difficuity of the Lycopodineous phylogeny, by making
clear that the foliar fibrovascular supply in the case of the Abietineae is of the same double
type as that which is primitive for the Lyginodendreae, the Cordaitales, the Cycadales,
the Ginkgoales, and many Filicales.

SUMMARY.

1. The Abietineae are divisible, on the evidence supplied by a study of their vegeta-
tive and reproductive organs, into two very distinct subfamilies, viz., the Pineae and the
Abieteae.

2. The Pineae are characterized by the invariable presence of resin canals, forming
an anastomosing system in the secondary wood and cortex of root and shoot. Resin
canals are present in the outer margin of the primary wood of the root. The scales of
the female cone are not deciduous. Pinus, Picea, Larix, Pseudotsuga.

3. The Abieteae ordinarily do not possess resin canals in the secondary wood of
root and shoot. Resin canals, however, are sometimes found in the wood of the female
reproductive axis, and in the first annual ring of vigorous shoots of sexually mature trees.
Resin canals occur in the secondary wood in tangential rows, as a result of injury. Resin
canals are invariably found in the center of the primary wood of the root. The scales of
the female cone are generally deciduous. Abies, Pseudolarix, Cedrus, Tsuga.

4. The evidence derived from anatomy and experimental morphology goes to show
that the presence of resin ducts in the woody tissues and in the cortex of the Abietineae
is primitive for the group. The resin canals persist longest in the reproductive axis, the
leaf, and the first annual ring of root and shoot. In the more specialized genera
the resin canals of the wood are replaced by resin cells, but in the latter condition

ANATOMY AND PHYLOGENY OF THE CONIFERALES. 33

of the wood, resin canals may always be recalled as a result of injury. The disappearance
of resin canals and their replacement by resin cells is probably for the sake of economy
of carbo-hydrate material. In Pseudolarix and Tsuga even the cortical resin canals
disappear from all organs except the female reproductive axis, together with its append-
ages, and the vegetative leaf.

5. The Abietineae are an older group than the Cupressineae, in the larger sense,
and are either antecedent to these or from the same ancestry. This conclusion is reached
from an anatomical and experimental morphological study of their organs, root, shoot, and
leaf. It is confirmed by the examination of the female reproductive organs and of the
pollen. It is further in conformity with palaeontological evidence.

6. The Abietineae are throughout characterized by the same double leaf trace which
is a constant feature of the older Gymnosperms, the Lyginodendreae, the Cordaitales, the
Ginkgoales, and the Cycadales. This feature serves to separate them from the Cupressi-
neae in the larger sense, and to unite them with the Cordaitales, which they resemble in
other important particulars, described in the body of the memoir.

7. The Abietineae must be regarded on comparative anatomical and morphological
grounds as a very ancient order of the Coniferales, and may even be the oldest living rep-
resentatives of this group.

In conclusion, I wish to express my obligation to Prof. G. L. Goodale for access to
the specimens of wood in the economic collections of the Museum. I am indebted to my
father for material collected during his summer vacation. My thanks are also due to Mr.
George S. Shaw, former chairman of the visiting committee to the Museum, and through
him to Prof. C. S. Sargent and Mr. Jackson Dawson of the Arnold arboretum of Harvard
university, and to Mr. Walter Hunnewell, the present owner of the Hunnewell pinetum.
But for the kindness of Mr. Shaw and those he has interested in my needs, this work
could scarcely have been carried to a conclusion. A considerable amount of valuable
material has reached me from the authorities of the Royal gardens, Kew, England, for

which I likewise wish to offer my sincere thanks.

PHANEROGAMIC LABORATORIES
or Harvarp UNIVERSITY.

34 EDWARD C. JEFFREY ON

LITERATURE.

Anderson, A. P.
°97. Comparative anatomy of the normal and diseased organs of Abies balsamea affected with Aecidium elatinum.
Bot. gaz., vol. 24, p. 309-344, pl. 14-15.
Bary, A. De.
*84. Comparative anatomy of the vegetative organs of the Phanerogams and ferns. Oxford: xvi + 659 pp., illus.
Bertrand, C. E.
"74. Anatomie comparée des tiges et des feuilles chez les Gnétacées et les Coniféres. Ann, sci. nat., bot., ser. 5, vol.
20, p. 5-128.
Coulter, J. M , and Chamberlain, C. J.
:O1. Seed plants. Vol.1. Gymnosperms. New York: 185 pp.
Cushman, J. A.
:03. Studies of localized stages in some plants of the Botanic gardens of Harvard university. Amer. nat., vol. 87,
p. 243-259.
Dawson, J. W.
°88. The geological history of plants. New York: 290 pp., illus.
Faull, J. H.
:O1. The anatomy of the Osmundaceae. Bot. gaz., vol. 32, p. 881-420, pl. 14-17.
Goebel, K.
°89-’93. Pflanzenbiologische schilderungen. Marburg.
Jackson, R. T.
°99. Localized stages in development in plants and animals. Mem. Boston soc. nat. hist., vol. 5, no. 4, p. 89-153, pl.
16-25.
Jeffrey, E. C.
°99. The development, structure, and affinities of the genus Equisetum. Mem. Boston soc. nat. hist., vol. 5, no. 5, p.
155-190, pl. 26-30,
:02. The structure and development of the stem in the Pteridophyta and Gymnosperms, Phil. trans. roy. soe. London,
ser. B, vol. 195, p. 119-146, pl. 1-6.
:03. The comparative anatomy and phylogeny of the Coniferales. Part 1.— The genus Sequoia. Mem. Boston soc.
nat. hist., vol. 5, no. 10, p. 441-459, pl. 68-71.
Kent, A. H. (ed.)
:00. Manual of the Coniferae; containing a general review of the order, a synopsis of the species cultivated in Great

Britain, their botanical history, economic properties, place and use in arboriculture, ete. Chelsea, [Eng]: 562 pp.,
illus.; James Veitch and sons, publishers.
Kidston, R.
:03. On the fructification of Neuropteris heterophylla Brongniart. Proc. roy. soc. London, vol. 72, p. 487.
Mayr, H.
°84. Entstehung und vertheilung der secretions-organe der fichte und lirche. Eine vergleichend-anatomische studie.
Bot. centralbl., vol. 20, p. 23-25, 53-56, 86-91, 117-121, 148-1538, 183-190, 213-219, 246-258, 278-288, 308-310, pl. 1-3.
Oliver, F. W., and Scott, D. H.
:03. On Lagenostoma lomaxi, the seed of Lyginodendron. Proc. roy. soc. London, vol. 71, p. 477-481.
Penhallow, D. P.
°96. The generic characters of the North American Taxaceae and Coniferae, Proc, and trans. roy. soc. Canada, ser,
2, vol. 2, sec. 4, p. 38-57, pl. 1-6,

Oe ——— a

ANATOMY AND PHYLOGENY OF THE CONIFERALES. 35

“:00. Notes on the North American species of Dadoxylon, with special reference to type material in the collections of
the Peter Redpath museum, McGill college. Proc. and trans. roy. soc. Canada, ser. 2, vol. 6, sec. 4, p. 51-97, 9 pls.
:03. Notes on Tertiary plants. Proc. and trans. roy. soc, Canada, ser. 2, vol. 9, sec. 4, p. 83-71, 12 pls.
Potonie, H.
*99, Lehrbuch der pflanzenpalaeontologie mit besonderer riicksicht auf die bediirfnisse des geologen. Berlin.
Radais, M.
94. Contribution a l’étude de l’anatomie comparée du fruit des Coniféres. Ann. sci. nat., bot., ser. 7, vol. 19, p. 165-
868, pl. 1-15.
Scott, D. H.
’96. Present position of morphological botany. British assoc. adv. sci., trans. of section K, botany, 1896, p. 1-19.
’97. The anatomical characters presented by the peduncle of Cycadaceae. Annals of bot., vol. 11, p. 899-419, pl.
20-21.
98. Onthe structure and affinities of fossil plants from the palaeozoic rocks.— II. On Spencerites, a new genus of
Lycopodiaceous cones from the coal measures, founded on the Lepidodendron spenceri of Williamson. Phil. trans.
roy. soc. London, ser. B, vol. 187, p. 83-106, pl. 12-15.
:00. Studies in fossil botany. London: 583 pp., illus.
:02. On the primary structure of certain palaeozoic stems with the Dadoxylon type of wood. Trans. roy. soc. Edin-
burgh, vol. 40, pt. 2, p. 331-365, 6 pls.
Solms-Laubach, H.
°90. Die sprossfolge der Stangeria und der iibrigen Cycadeen. Bot. zeitung, vol. 48, p. 177-187, 193-199. 209-215,
225-280, pl. 2.
Stopes, M. C.
:03. On the leaf structure of Cordaites. New phytologist, vol. 2, p. 91-98, pl. 9.
Strasburger, E.
~ 172. Die Coniferen und die Gnetaceen, Eine morphologische studie. Jena: 8°, with atlas of 26 pls., 4°.
‘91, Histologische beitriige. 3. Ueber den bau und die verrichtungen der leitungsbahnen in den pflanzen. Jena:
xxxii + 1000 pp., illus.
Thomas, F.
766. Zur vergleichenden anatomie der Coniferen-laubblatter. Pringsheim’s jahrb. f. wissensch. bot., vol. 4, p. 23-60.
Williamson, W. C.
*81. On the organization of the fossil plants of the coal measures.— Part 11. Phil. trans. roy. soc. London, vol. 172,
pt. 2, p. 283-805, pl. 47-54.
Williamson, W. C., and Scott, D. H.
795. Further observations on the organization of the fossil plants of the coal-measures.— Part 38. Lyginodendron and
Heterangium. Phil. trans. roy. soc. London, ser. B, vol. 186, p. 7038-779, pl. 18-29.
Worsdell, W. C.
197. On “transfusion-tissue ”: its origin and function in the leaves of Gymnospermous plants. Trans. Linn. soc. Lon-
don, ser. 2, botany, vol. 5, pt. 8, p. 801-319, pl. 23-26.
Zeiller, R.
:00. Eléments de paléobotanique. Paris: 421 pp., illus.
:04. Observations au sujet du mode de fructification des Cycadofilicinées. Comptes rendus de l’acad. des sci., Paris,
vol. 188, no. 11, p. 663-665.
Zeiller, R., and Fliche, P.
:03. Découverte de strobiles de Sequoia et de Pin dans le Portlandien des environs de Boulogne-sur-Mer. Comptes
rendus de l’acad. des sci., Paris, vol. 187, p. 1020-1024.

36 EDWARD C. JEFFREY ON

EXPLANATION OF PLATES.

°

(All the figures are from photomicrographs.)

PLATE 1.

Fig. 1. Transverse section through the margin of the traumatic wood in Pseudotsuga taxtfolia. X 20.

Fig. 2. Transverse section of the root of P. taxifolia. X 20.

Fig. 3. Transverse section of the lower third of the leaf of P. taxifolia. X 40.

Fig. 4. Transverse section of a stem of P. taxifolia, showing the exit of the leaf trace from the central cylinder. X 60

Fig. 5. Transverse section of the stem of P. taxifolia, showing the structure of the leaf trace in its passage through
the cortex. X 60.

Fig. 6. Transverse section of the root of Larix americana, showing the communication of the resin ducts, opposite
the groups of protoxylem, with the ducts of the cortex. X 60.

Fig. 7. Transverse section through the leaf base of L. americana, showing the double trace. X 60.

Fig. 8. Transverse section through part of a wounded stem of Picea nigra, showing the formation of traumatic resin

eanals. X 10.

PLATE 2.

Fig. 9. Transverse section of the stem of Picea nigra, showing the double leaf trace. X 60.

Fig. 10. Transverse section of a wounded stem of Abies balsamea, showing the formation of traumatic resin canals
along the margin of the wood. X 8.

Fig. 11. Part of the same more highly magnified to show the details of the resin canals. X 30.

Fig. 12. Traumatic wood of the base of a witches’ broom of A. balsamea, showing the formation of resin canals.
x 380.

Fig. 13. Section of the injured wood of the stem of A. magnifica, showing the formation of numerous rows of resin
canals in the same annual ring. » 60.

Fig. 14. Transverse section through the upper part of the axis of the cone of A. magnifica, showing the presence of
normal resin canals in this organ. X 20.

Fig. 15. Section through part of the lower region of the cone-axis of A. magnifica, showing the smaller size of the
resin canals. XX 30.

Fig. 16. Section through part of the same region as that shown in Fig. 14 more highly magnified to show the fusion of
the resin canals. XX 40.

PLATE 3.

Fig. 17. Section through a vigorous branch in the reproductive region of Abies magnifica, showing the presence of
normal resin canals in the first annual ring. X 20.

Fig. 18. Section through another specimen more highly magnified to show the normal occurrence of resin canals in
the first annual ring. X 20.

Fig. 19. Section through a vigorous yearling branch of A. magnifica, showing the presence of normal resin canals.
x 20.

Fig. 20. Transverse section of the inner portion of the last, highly magnified to show the structure of the resin canals.
x 180.

Fig. 21. Longitudinal section of the pith of A. magnifica, showing the presence of sclerified diaphragms similar to
those of the Cordaites. » 25.

ANATOMY AND PHYLOGENY OF THE CONIFERALES. 37

Fig. 22. Transverse section of a root of A. nobilis, showing the median resin duct of the primary wood. 15.

Fig. 23. ‘Transverse section of the leaf trace of A. magnifica, showing the jacket of transfusion tissue surrounding the
phloem. > 150.
Fig. 24. Transverse section of the cone-axis of A. appolinis showing the presence of normal resin canals. X 20.

Fig.
Fig.
Fig.
Fig.
Fig.
Fig.
Fig.
Fig.

Fig.

canals.

Fig.
Fig,
Fig.
Fig.
Fig.
Fig.
Fig.

X 20.

Fig.
Fig.
Fig.
Fig.

x 20.

Fig.
Fig.
Fig.
Fig.

Fig,
Fig.
Fig.
Fig.
Fig.
Fig.
Fig.
Fig.

PLATE 4.

25. Part of the last more highly magnified. » 40.
26. Section of a wounded branch of Pseudolarix. x 8.
27. Part of the same more highly magnified, to show the presence of traumatic resin ducts. 40.
28. Section of the root of Pseudolarix, to show the median resin canal. »X 20.
29. Part of another root more highly magnified. x 80.
80. A wounded root of Pseudolarix. X 20.
31. Transverse section of a branch of Pseudolarix. » 15.
82. Transverse section of the peduncle of a female cone of Pseudolarix. X 20.
PLATE 5.
33. Section of a portion of the axis of the female cone of Pseudolarix, showing the presence of cortical resin
x 80. .
34. Section through the base of the ovuliferous scale and the sterile bract of Pseudolarix. » 25.
35. Section through the sterile bract of the same. X 40.
86. Section through the foliage leaf of Pseudolarix. »X 40.
37. Section through traumatic wood of Cedrus atlantica. X 20.
38. Section through the wounded root of the same. X 20.
89. Resin canals of the same more highly magnified. > 180.
40. Section through the wood of C. deodara, showing the presence of both horizontal and vertical resin canals.
PLATE 6.
41. Another section more highly magnified. x 70.
42. Longitudinal section of wood of the same, showing horizontal and vertical resin canals. X 15.
43, Another of the same. X 43.
44. A tangential section of the same, showing resin canals of various degrees of development in transverse section.
45. Section of cone of Cedrus atlantica, to show absence of resin ducts in the wood. X 30.
46. Transverse section of the root of C. atlantica, to show median resin canal. X 20.
47. Transverse section of the double cortical leaf trace of C. atlantica. X 120.
48. Transverse section of wounded stem of Tsuga canadensis. X 10.
PLATE 7.
49. Transverse section of a vigorous yearling stem of Tsuga mertensiana. X 20.
50. Transverse section through part of the cone of thesame. x 25.
51. ‘Transverse section of the axis of the cone, T. canadensis. 25.
52. Transverse section of root of T. mertensiana. X 20.
53. Transverse section of root of T. canadensis. X 30.
54. Transverse section of part of a wounded root of T. canadensis. x 30.
55. Transverse section of center of root of the same. X 180.
56. Transverse section of stem of the same showing double leaf trace. x 40.

Printed January, 1905.

MEMOIRS BOSTON Soc. NAT. HIST. VOL. 6.

SS

—=

Mf
oe
W

JEFFREY ON THE ABIETINEAE.

bo, as

MEMOIRS BOSTON Soc. NAT. HIST. VOL. 6.

\
a
at

own

> id a

-

of

JEFFREY ON THE ABIETINEAE.

. |
’

MEMOIRS BOSTON Soc. NAT. HIST. VOL. 6. PLATE 3

23 24
JEFFREY ON THE ABIETINEAE.

MEMOIRS BOSTON Soc. NAT. HIST. VOL. 6, TE 4

JEFFREY ON THE ABIETINEAE.

&

VOL. 6.

BOSTON Soc. NAT. HIST.

|

>
\

MEMOIR:

40

39

JEFFREY ON THE ABIETINEAE.

“~

MEMOIRS BOSTON Soc. NAT. HIST. VOL. 6. PI

To

sise~,|)\\eagcelaseiges | \
si) aM \ | ee PAM TM J
Men) / (eee niet : f A 3

ey

U
Ney

J
.

ene
98

tt
ey
-
ra
ee
. id
&
@
4
ud

JEFFREY ON THE ABIETINEAE.

MEMOIRS BOSTON Soc. NAT. HIST. VOL. 6. PLATE 7

a eh
Sate
See Ms

eg

JEFFREY ON THE ABIETINEAE.

¥
, A ‘ %
F oe) PP ae ye Jae
7
ue
ie"
7
i oe :
aS eg
ae :
i
a Li ry q
. 7 > 3 i
ar] ’ ‘Cy 7 A
>, e /
we ’ a ee
tl
ae | |
bg ; a >
‘ 4
5 .
* : ’ |
r tah
v 4 : *
F w
: wt os
F ; aide:
J i
« ie Die ‘

ie
ia
nie

Parry

Ths?
wha he
i Dar,

,
Pehl)
f
t f
‘ -
oT y
‘a
ry i
, 7
. t 1 Lie oll
ay age >
+ : aq
’ is =
d a -
’ nef ers
as
F
i if
" yt ;
oF t . » >
‘ : d 7
ss re) a
ios 4 AD ty “ " 1 i - ,
he ey "at * +). 3S $e 4 =a
¥, Ee ams Le. . }
A Pas et hs 8 lhe
-" an 5
a . z

sei

JE

7.