Carnegie Institution of Washington publication

IBPi

IP
mm

¥m

BS

THE BROAD-SCLEROPHYLL VEGETATION
OF CALIFORNIA

AN ECOLOGICAL STUDY OF THE CHAPARRAL AND
ITS RELATED COMMUNITIES

BY

WILLIAM S. COOPER

Published by the Carnegie Institution of Washington
Washington, October, 1922

CARNEGIE INSTITUTION OF WASHINGTON
Publication No. 319

1 6 192

TECHNICAL PRESS
WASHINGTON, D. C.

CONTENTS.

Page.

Introduction °

Relations of the chaparral to other
communities of scrub in North

America "

Previous work '

Scope of present paper 9

Acknowledgments 10

I. Range and center of distribution. ... 11
Location of greatest differentiation

of type I 2

Location of dominance or great

abundance of individuals 14

Location of synthetic or closely re-
lated forms ■ • • • 14

Location of maximum size of indi-
viduals IS

Location of least dependence upon

a restricted habitat 15

Location of greatest importance
of the type in the climax commu-
nity 16

Continuity and convergence of

lines of dispersal 16

II. Climatic relations 17

III. The communities 20

The broad-sclerophyll communities 20
The broad-sclerophyll forest for-
mation 21

The chaparral formation 25

The climax chaparral associa-
tion 26

The conifer forest chaparral

association 27

IV. Vegetation and habitat 30

The locality for intensive study. . . 30

The Palo Alto region 30

Jasper Ridge 30

The vegetation 32

The habitat 42

The soil factors 42

Page.
IV. Vegetation and habitat— Cont'd.

Physical character 42

The soil moisture 43

The soil temperature 55

The atmospheric factors 56

Light 56

Temperature 57

Wind 58

The atmospheric moisture .... 58

Discussion and correlations 63

The broad-sclerophylls and their

habitat 63

The habitats of the two climaxes

contrasted 67

V. Development 72

The climax regions of California . . 72

The broad-sclerophyll climax forest 73

The climax chaparral 75

Evidence of its climax character 75
Extent of the chaparral as a true

climax "6

Successions leading to the chap-
arral climax 82

Primary successions 82

Secondary successions 86

VI. Ecological character of the broad-

sclerophyll shrubs and trees .... 88

Growth form 88

Root system 89

The leaf 92

The deciduous element 92

External characters 93

Internal structure • 97

Effects of environmental condi-
tions upon leaf structure 107

Comparative transpiration-rate . 1 10
Appendix: Annotated list of broad-
sclerophylls and accompanying spe-
cies 113

Bibliography 122

I 7 H 3

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

By William S. Cooper.

INTRODUCTION.

The characteristic vegetation of California west of the high
Sierra Nevada and the Colorado and Mojave Deserts is of the type
aptly described by the term "broad-sclerophyll." The use of the
prefix "broad," which expresses a very general group character,
in distinction to the needle-leaf of the "narrow-sclerophyll" conifers,
is not without its disadvantages, since the most important chaparral
species, Adenostoma fasciculatum, possesses a needle-like leaf. This
single though important exception seems not to be sufficient cause
for discarding the very expressive appellation.

Broad-sclerophyll vegetation is not confined to California, but
recurs upon other portions of the earth's surface, notably the shores
of the Mediterranean. The leaf character is the conspicuous and
diagnostic feature, that organ being thick, stiff, and hard, ordinarily
flat, and evergreen. Schimper (80) has shown that this vegetation
type is everywhere correlated with a definite type of climate, namely,
one with a long, dry summer and a rainy winter. Some of the broad-
sclerophylls are trees, but most are scrubs. We therefore find
broad-sclerophyll forest and broad-sclerophyll scrub, the latter
being the more widespread and important. The present work is
an ecological study of the broad-sclerophylls of California; of their
relations to climate and soil and to each other. Being somewhat
of a pioneer work, many phases of the problem are touched, and
many lines of investigation have been opened up which could not
be followed to the end. The field is a fascinating one, and a lifetime
would not suffice to exhaust its possibilities.

In the Mediterranean region broad-sclerophyll scrub is known as
"macchie" and "garigue"; in California it is called "chaparral." It
has already been noted that the scrub is more important than the forest,
both scientifically and economically. In fact, the present research
began as an investigation of the chaparral alone. Because of the close
relations between them, it was a simple matter to extend the field of
study to include the trees. The term "chaparral" is of Spanish origin,
being derived from the word "chaparra," meaning scrub oak. It
seems to have been applied by the early explorers of California to
the low, shrubby, dominantly evergreen vegetation which they found
to be so characteristic of the Coast Ranges and the foothills of the
Sierras. Locally the term is sometimes restricted to a single species,
often Ceanothus cuneatus. The term ' ' chamisal ' ' is frequently applied
to a pure growth of Adenostoma fasciculatum or "chamise."

6 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

RELATIONS OF THE CHAPARRAL TO OTHER COMMUNITIES
OF SCRUB IN NORTH AMERICA.

It will be well at this point to differentiate between the chaparral
and the other scrub communities of western North America. The four
great types may be characterized as follows, no attempt being made to
describe them minutely or to relate them to environmental conditions.

The sagebrush type. — The characteristic species, Artemisia triden-
tata, always makes a large part and frequently grows in purity over
immense stretches. This is preeminently the type of the Great
Basin, though extending beyond it more or less in all directions.
It is sharply differentiated from the chaparral by reason of its abso-
lutely different leaf-type and by the fact that the two overlap very
slightly in range.

The desert scrub.— This is made up of a multitude of species not
at all uniform superficially in ecological character. The succulents
are very important, opuntias of both the cylindrical and flat-jointed
forms being especially prominent. There are also thorny shrubs,
some of them with soft deciduous leaves. Others, with very small
leaves, have been well named microphylls. Finally, a few species
in aspect recall the chaparral, the most important of these being
Comllea tridentata. The desert scrub ranges from Arizona to Texas
and southward into Mexico. In ecological character it is sharply
set off from the chaparral, except for the superficial resemblance
of Covillea just noted. This species, although it ranges to the very
border of the chaparral country, mingles with that type practically
not at all. Cannon (20), moreover, states that "up to this time
.... all attempts to grow it at the Coastal Laboratory [Carmel,
Monterey County] have failed." He explains this on the basis of
less efficient soil temperature, but the value of that factor is of course
a direct result of climatic causes. There is thus some element in
the ecological make-up of the species which unfits it for life in the
chaparral region, in spite of its superficial likeness. Here is a warning
against the placing of too much reliance in the working out of ecologi-
cal relationships upon a single conspicuous character, or even upon
structural characters in general. Vital processes rather than struc-
ture are fundamental.

The deciduous thicket type. — Most important here are scrub oaks
of various species, and the type is a widespread one, occurring from
the Atlantic to the Pacific and from Canada to Mexico. Great
stretches of foothills in the central Rockies are densely covered by it.

This is the closest of the three to the chaparral, both floristically
and ecologically. In contrast to the first two, the transition between
this type and the chaparral is gradual. In the Sierras and in northern
California much of the scrub, dominantly evergreen, has nevertheless
a large deciduous element, and there are considerable thicket areas
which are composed entirely of deciduous species.

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 7

A few of the Californian sclerophylls, too, extend far beyond the
average limits of the type. Arctostaphylos pungens ranges eastward
to southwestern Colorado, where it forms an unimportant part of
the deciduous thickets. Moreover, Quercus undulata, an important
member of the Rocky Mountain thickets, itself tends strongly
toward evergreenness.

The chaparral— The chaparral may be denned as a scrub com-
munity, dominated by many species belonging to genera unrelated
taxonomically, but of a single constant ecological type, the most
important features of which are the root system, extensive in pro-
portion to the size of the plant, the dense rigid branching, and pre-
eminently the leaf, which is small, thick, heavily cutinized, and
evergreen. This definition might be applied with equal accuracy
to the macchie of the Mediterranean regions; chaparral and macchie
appear to be ecologically equivalent. The chaparral is characteris-
tically, almost exclusively, of California west of the Sierra crest and
the deserts; in other words, of the region of "Californian" climate.

Because of its ecological distinctness, the Californian broad-
sclerophyll scrub is entitled to a name of its own, and the term
"chaparral," by reason of almost universal usage, is the obvious
choice. The desert scrub and the deciduous thicket, which have
been called chaparral by various authors, can easily be provided
with other names if the ones used here are not satisfactory. Finally,
it should be emphasized that the chaparral finds its closest ecological
relative not in any other scrub community, but in the broad-sclero-
phyll forest, which is a response to the same type of climate.

PREVIOUS WORK.

Practically no work of purely ecological nature dealing with the
chaparral has been published up to the present time. Several
foresters have written excellent accounts of the aspect and behavior
of the brush, especially in its relation to the forests which are asso-
ciated with it. This emphasis is natural, since the chaparral is
extremely important to the forester whose field of work is in Cali-
fornia. It is essential for him to determine whether the brush can
be replaced by a more valuable crop and, if not, how it can best be
made to perform its important economic function of water-shed
protection. In addition, there have been taxonomic studies of
certain genera, and phytogeographic researches of more or less
general nature, in which the chaparral has been treated. The
broad-sclerophyll forest has been almost wholly neglected, except in
its taxonomic aspects. It will be convenient to arrange the following
brief survey of the literature under four heads, adding a miscellaneous
group for certain papers that can not be classified under the three
mentioned above.

(a) Detailed notices of taxonomic researches are outside the

8 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

province of this work. Four papers dealing with difficult genera
should be mentioned: Abrams on Uva-Ursi [Arctostaphylos] (2),
Trelease and Brandegee on Ceanothus (89, 10), and Miss Eastwood
on Garry a (28).

(6) Papers by foresters may be grouped in two classes: (1) those
dealing particularly with chaparral; (2) descriptions of various
forest reserves in which chaparral occurs. One of the former,
by Plummer (74), purports to be a rather full discussion of the subject
from both the purely scientific and the economic standpoints.
It is totally inadequate. All the other papers by foresters deal
with conditions in more or less restricted localities. Those by Boerker
(9), Foster (30), Haefner (37), and Sterling (85) treat of northern
California, where the chaparral is mainly of the temporary kind.
Naturally the successional relations with the forest trees and the
effects of fire upon these are treated with considerable detail. Sub-
division of the general type into communities and subcommunities
is attempted to some extent, and also the correlation of these in a
very general way with habitat factors, especially slope exposure.
The work has been almost wholly observational, with no exact
habitat measurements and few statistical studies of the vegetation.
Miller (64), studying the chaparral of northern California as a
watershed cover, gives some data bearing upon successional relations.
In the works of more general scope by Barber (6), Leiberg (50-56),
and Sudworth (86) there is considerable material of value relating
to the extent, composition, zonation, and successional relations
of the chaparral in various parts of the State. E. N. Munns has put
forth three papers (66-68), in which some instrumental habitat
data are presented. He has also much valuable material, as yet
unpublished, to which I have had access through his kindness.

(c) Four papers dealing with the phytogeography of southern
California must be mentioned. McKenny (60) recognizes in Orange
County seven formations, two of which (the mountain and the
foothill formations) are mainly chaparral of various types, which
are not very clearly distinguished. Hall, in his account of the San
Jacinto Mountains (38), relates the vegetation to the well-known
temperature zones of Merriam, and distinguishes two altitudinal
belts of chaparral, the lower dominated by Adenostoma, the upper
by Castanopsis and other genera. He also summarizes the obvious
ecological characteristics of the average chaparral shrub, noting
reduction of leaf-surface, vertical position, thickness and leathery
quality, frequency of woolly pubescence (with which I would dis-
agree), and depth of penetration of root system. Parish (71)
divides southern California into five phytogeographic areas, and in
consideration of the cismontane area describes the chaparral in
general terms. Abrams (1) emphasizes the facts of zonation, ad-
hering to the Merriam arrangement, and distinguishes three altitudi-

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 9

nal belts of chaparral: the lowest, dominated by half-shrubs such
as Ramona and Eriogonum (not to be classed as chaparral in my
opinion); the second, in which Adenostoma is most important; and
the upper, in which Arctostaphylos and other genera are in control.

The more recent phytogeographic handbooks contain brief descrip-
tions of Calif ornian vegetation, with more or less accurate references
to the chaparral. Drude (2G) applies the term to the very different
desert scrub of western Texas, southern New Mexico, and northern
Mexico, and has little to offer upon the sclerophyll vegetation of
California. Engler's treatment (29) is similar. Schimper (80) is
the first of the European geographers to give an adequate description
and interpretation of the California sclerophyll vegetation. His
information is mostly obtained apparently from a paper by Purpus
(76). He gives a brief but adequate description of the chaparral,
its general character, leaf-type, and relation to climate, and shows its
likeness to the vegetation of other regions with winter rains. Warm-
ing (90) uses the term chaparral in the same sense as Drude does,
but recognizes the Californian type and its relation to similar ones
in other regions in the sentence : " In California maqui is known under
the name of chaparral." Harshberger's account (41) is merely a
compilation from Drude, Jepson, Parish, and McKenney. Clements
(22) gives a brief description of the California scrub, treating it as
the third of three subdivisions of the Chaparral Formation, which
he conceives very broadly: the Petran (Rocky Mountain), the
Sub-climax, and the Coastal Chaparral. The first two correspond
with the "deciduous thicket" of the present work.

(d) Among miscellaneous papers should be mentioned a local
descriptive article by Purdy (75); three by Cannon — one (17)
showing the dependence of the Monterey pine (Pinus radiata)
upon prevention of soil desiccation by accompanying chaparral,
and two (18, 19) in which the relation of Quercus agrifolia and other
trees to soil-moisture is treated; one by Brandegee (12) describing
the recovery of the chaparral after fires; one of similar scope by Jep-
son (49) ; and an account of the distribution of species of Eriodictyon
by Abrams and Smiley (3). Incidental references to various phases
of the ecology of the chaparral are found in the publications listed
in the Bibliography as Nos. 24, 27, 33, 34, 35, 44, 45, 46, 61, 62.

SCOPE OF PRESENT PAPER.

It will be seen that investigation has progressed to a more or less
satisfactory degree along the following lines: taxonomy, ranges of
species, range of type, its subdivisions, relation to similar types in
other regions, to climate, and to fire. The ecological character
of the different species has been barely touched; the same is true
of the mutual relations of vegetation and habitat, and the develop-
mental phases of the problem have been treated only so far as fire

10 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

is a factor in them. These subjects are of great scientific interest
in themselves. They are also fundamental to a proper understanding
of the economic relations of the species. The main emphasis in
this paper is laid upon these three subjects, though new material
is offered along all the other lines mentioned, except taxonomy.
The plan of treatment is to pass from the general to the specific.
The first two chapters deal with phytogeographic relations, the
next three with local units, their development and relation to habitat,
and the last treats of the ecological character of the individual plant.
An annotated list, with much detail that may not interest the average
reader, has been added as an appendix.

The field work was of the two usual sorts: intensive instrumental
and quadrat study, and extensive exploration. The former was
carried on mainly in the vicinity of Palo Alto from November
1912 to September 1915, and in the summers of 1916, 1917, and
1919. The exploration covered the whole extent of the Coast Ranges
from San Diego to Eureka, and a number of representative localities
in the Sierras.

ACKNOWLEDGMENTS.

I wish to express first my deep appreciation of the very great
kindness of every one of the faculty of botany at Leland Stanford
Junior University, who put all the facilities of the department at
my disposal and who frequently gave me their personal assistance
in many ways. Mr. S. B. Parish of San Bernardino, Dr. L. R.
Abrams of Stanford University, and Dr. H. M. Hall and Dr. W. L.
Jepson of the University of California read the manuscript of the
floristic portion and made important corrections and suggestions.
Dr. Charles F. Shaw, of the University of California, furnished
the mechanical analyses of the soils from Jasper Ridge, and Dr.
Charles B. Lipman supplied the humus determinations of the same
samples. Dr. D. T. MacDougal, of the Carnegie Institution of
Washington, offered valuable advice and encouragement, and
the Carnegie Institution rendered financial assistance which was
applied to the purchase of apparatus. Mr. Frank Shaw, of Redwood
City, gave faithful and efficient assistance in the field work at Palo
Alto. The photo-micrographic negatives for plate 20 were the
work of Dr. C. O. Rosendahl, of the University of Minnesota;
Miss Vinnie A. Pease executed the camera drawings of the leaf-
sections (figs. 22-43), and Miss Elsa Horn made the multitudinous
measurements and calculations involved in the studies of leaf char-
acter in relation to habitat and of transpiration. Miss Horn also
rendered efficient service in the carrying out of experimental soil
studies and in making certain drawings. To all these persons I
wish to express my sincere gratitude.

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 11

I. RANGE AND CENTER OF DISTRIBUTION.

The term "center of distribution" may be understood in two
ways. It may mean, on the one hand, the region in which one or
more species have originated and from which they have spread;
or it may signify the region where a species or a group of species
attains its greatest development. The present study is not primarily
concerned with origins and migrations, and therefore the second
interpretation is the one used. It is true, however, that in most
cases, and probably in the one here considered, a given center is a
" center of distribution" in both senses. Transeau, in an important
paper (88), has expressed the concept as follows:

"In using the term 'center of distribution' it is not implied that the plants have
necessarily spread from these centers, but that the complex of climatic factors most
favorable to the development of this type of vegetation is here localized, and that as
we depart from such centers we find conditions more and more unfavorable."

Referring to a particular center he says that "within its limits,
the plants have a wider range of habitats, attain a greater size,
and are more abundant than elsewhere."

Adams (5) gives a list of ten criteria for the determination of
centers of distribution, which include the three stated by Transeau.
Six of these are susceptible of use in the present study. They
are as follows:

1. Location of greatest differentiation of a type.

2. Location of dominance or great abundance of individuals.

3. Location of synthetic or closely related forms.

4. Location of maximum size of individuals.

5. Location of least dependence upon a restricted habitat.

6. Continuity and convergence of lines of dispersal.

To these I would add another, stated as follows: Location of
greatest importance of the type in the climax community. There
is evidence along all of these lines in support of the conclusion about
to be stated. Another criterion proposed by Adams, and which
would be of great use were our data sufficient, is worth stating, as a
suggestion for future study: "Continuity and directness of indi-
vidual variations or modifications radiating from the center of
origin along the highways of dispersal."

From the evidence about to be presented I have reached the con-
clusion that the center of distribution of the broad-sclerophyll
vegetation type of the Pacific Coast corresponds with the region
known popularly as "Southern California," west of the deserts,
including Ventura and Santa Barbara Counties, and also extending
an unknown distance into Lower California. The evidence can
best be stated under seven heads, corresponding with the six criteria
proposed by Adams and the seventh which I have added. At thi?

12 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

point I intend only to establish the location of the center. The
environmental phases of the problem will be considered later.

LOCATION OF GREATEST DIFFERENTIATION OF TYPE.

Our first problem under this head is to discover which part of the
general region inhabited by the type is richest in species of that type.
As a corollary to this, the relative number of endemics in the different
areas may be compared. The former is easily done by plotting the
ranges of all the species concerned upon a single map and indicating
by density of shading the numbers of overlapping ranges. The
results are shown in plate 1, the large map in detail for the State of
California, the small one in more generalized form for the whole
region covered. Some practical difficulty was encountered in super-
posing the ranges of so large a number of species. The problem
was solved by cutting the ranges from thin cardboard, gluing them
together in their proper positions, outlining the total extent indicated,
cutting away all portions where five thicknesses of cardboard or less
occurred, outlining the remainder, and repeating the process. It is
scarcely necessary to state that the map must be interpreted broadly
and that details must be disregarded, since in only a few cases has it
been possible to draw the ranges with an approach to exactness.
The species considered are in general those of lists I and II (see
Appendix), the deciduous ones being excluded. A few of the broad-
sclerophylls were omitted because of lack of satisfactory knowledge
of their ranges. The total number of species considered is 71.

From the small map we see that, broadly speaking, the type has
its stronghold in California and Lower California. A few species
extend northward beyond the State line; about 21 into Oregon, many
of them a short distance only; 11 into Washington, and 8 into British
Columbia. None occurs north of California which does not enter
that State, unless we admit Howell's species of Arctostaphylos to
consideration. The complete list of those ranging northward beyond
California is as follows 1 :

Myrica californica.** Ceanothus cuneatus. Arbutus menziesii.**

Castanopsis chrysophylla.* sanguineus.** Arctostaphylos manzanita.

sempervirens. thyrsiflorus.* nevadensis.*

Quercus chrysolepis. velutinus.** patula.

sadleriana. Garrya elliptica. tomentosa.**

Pasania densiflora. fremontii. Gaultheria shallon.**

Umbellularia californica. Rhododendron californicum.** Vaccinium ovatum.**

Six species extend discontinuously eastward beyond the boundaries
of California:

Cercocarpus ledifolius. Ceanothus velutinus.

Rhamnus californica. Arctostaphylos patula.

crocea. pungens.

1 Unstarred species extend into Oregon only; those with a single star (*) into Washington,
and double-starred (**) species into British Columbia.

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 13

Over the great area of the Great Basin and the Rocky Mountains
covered by the lightest shading, the distribution of the few broad-
sclerophylls is of course not continuous. The extreme limits of
the ranges are indicated, detailed mapping in these cases being
impossible.

Focusing attention upon the State of California, the following
facts appear : First of all, it is seen that the Coast Ranges, the lower
and middle altitudes of the Sierras, and the coastal area of southern
California are the regions where the broad-sclerophylls are of im-
portance. In two places they are absent — the Great Valley and
the alpine regions of the Sierras. Over much of the former occasional
plants of a number of species occur, and the region might fairly
be covered by the lightest shading. This problem is discussed
later (p. 77).

The development of the type in the Sierras is moderate, the belt
of maximum number being in the region where the ranges of foothill
and forest species overlap.

The zone of deep shading in the Coast Ranges, southward through
southern California into Lower California, is the outstanding feature.
Northward this follows the inner ranges, but from Monterey Bay
south it borders the coast. Two spots of very special abundance
are seen, one extending from Marin County to the Santa Lucias,
the other covering the mountains of Santa Barbara and Ventura
Counties and the San Gabriel and San Bernardino Ranges. Too
much reliance should not be placed upon these, since both regions
have been explored and studied with comparative thoroughness.
It is true, however, that certain portions of these areas have been
proved to be rich in endemics, and it may well be, therefore, that
the seeming abundance of species in these spots is not wholly a false
appearance.

A division into regions, with lists of species confined to each, will
supplement the data furnished by the map:

In the Sierra Nevada 32 species occur, but 15 of these are found
also in all the other regions and only the 3 following are confined
to this region alone:

Ceanothus diversifolius. Ceanothus parvifolius. Arctostaphylos mariposa.

In the North Coast Ranges (San Francisco Bay being the point
of division), 38 species occur, and 1 is endemic, Arctostaphylos
stanfordiana. 1

In the South Coast Ranges 37 species occur, and the 6 following
are endemic:

Ceanothus dentatus. Arctostaphylos andersonii. Arctostaphylos pumila. 2

papillosus. 2 hookeri. 2 vestita. 2

1 Here, and in other lists, several species might be added if accurate knowledge concerning
their validity and range could be obtained.

2 Members of the group of endemics characteristic of the Monterey region, A. hookeri and
A. vestita being strictly confined to it.

14 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

In Southern California 47 species occur, and the following 13
are confined to the region or extend into Lower California:

Quercus engelmanni. Ceanothus crassifolius. Ceanothus oliganthus.

Adenostoma sparsifolium. megacarpus. Comarostaphylis diversifolia.

Rhus integrifolia. spinosus. Xylococcus bicolor.

laurina. verrucosus. Arctostaphylos drupacea.

ovata.

Summarizing from the map and the lists just given, we see first
that California and Lower California are the home of the broad-
sclerophylls. The number of species diminishes eastward very
rapidly, and northward along the coast more gradually. Further,
we see that the greatest number of species occur in the Coast Ranges
and Southern California. By a division into four areas we discover
that the Sierras are poorest both in total number of species and in
endemics; that the north and south Coast Ranges are next, being
about equal in total numbers, but very unequal in endemics, the
north portion having one and the south six; and that southern
California stands highest with respect to both categories. Our
evidence therefore points to southern California as the center of
distribution of the broad-sclerophylls, and confirmation of this
conclusion will be found in the paragraphs following.

LOCATION OF DOMINANCE OR GREAT ABUNDANCE OF

INDIVIDUALS.

It is a matter of easy observation that the chaparral fields have
their greatest extent and dominance in southern California. We
have only to cite as examples the mountains of Ventura County
(plate lie), the lower ridges of the San Gabriel, the San Bernardino
and San Jacinto Ranges, and the Cuyamaca Mountains (plate 11a).
Without doubt the same is true of northern Lower California.
Traveling northward, we find in both Coast Ranges and Sierras
a gradual decrease in the vegetational importance of the broad-
sclerophylls. The chaparral areas are more and more restricted,
and increasingly dominated by species of successional status. In
the southern Coast Ranges and Sierras the change is not so notable,
but it intensifies as we reach the central portion of the State. A
similar decrease is seen, very naturally, as we ascend to higher alti-
tudes in all parts of the region. In the northern portion of the
State and beyond its limits, the broad-sclerophylls lose their domi-
nance altogether and become less and less important vegetationally.
Eastward the decrease is very sharp, so that beyond the western
borders of the deserts and the middle forest region of the Sierras
the type is almost altogether absent as a vegetational entity.

LOCATION OF SYNTHETIC OR CLOSELY RELATED FORMS.
It is impossible, in the present state of our knowledge, to obtain
accurate data upon this point. However, the very confusion which

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 15

is so evident in the various taxonomic treatments of certain genera
is itself evidence after a fashion. The genera Ceanothus and Arcto-
staphylos are classic examples. In both, the number of species
proposed has been very large. Many of these fall into definite
groups having close relationship within themselves, and by some
authors these groups are treated as single variable species. A case
in point is the group-species known commonly as Arctostaphylos
tomentosa. From this Miss Eastwood has segregated three forms,
all of which grow upon Mount Tamalpais; Abrams refers all to
A. tomentosa. Howell has described 7 from southwestern Oregon,
which Abrams reduces to synonymy. In Ceanothus several groups
of closely related forms or species might be noted (Brandegee, 10),
and in Rhamnus and Garrya the structure is similar, but on a smaller
scale. It is evident that variation and hybridization are rampant
among certain genera of the California sclerophylls. It is impossible
to locate accurately the geographical area where these processes
are most active, further than to state that they are less evident in
the Sierras than in the Coast Ranges and southern California.

LOCATION OF MAXIMUM SIZE OF INDIVIDUALS.

In the case of the 9 broad-sclerophyll trees decision is easy; 4 of
them, Myrica californica, Quercus agrifolia, Q. chrysolepis, and
Q. wislizeni, attain their largest size in the central Coast Ranges,
and the last two (according to Sargent) also in the central Sierras;
4 more, Castanopsis chrysophylla, Pasania densiflora, U?nbellularia
californica, and Arbutus menziesii, make their best growth in north-
western California. Quercus engelmanni is confined to southern
and Lower California. As to the chaparral shrubs, data are scanty
and unsatisfactory, especially because fire so commonly brings the
shoots to an untimely end. As to the most important single species,
Adenostoma fasciculatum, the greatest average size that I have seen
was in San Benito County, in the south Coast Ranges. Other
species, ordinarily shrubs, attain respectable tree size in the same
general region, e. g., Heteromeles arbutifolia and Prunus ilicifolia
in the vicinity of Palo Alto. The very general statement may be
made that the broad-sclerophylls attain their greatest size in the
coastal region — the trees in the north and the shrubs in the south.

LOCATION OF LEAST DEPENDENCE UPON A RESTRICTED

HABITAT.

In southern California we find the broad-sclerophylls least con-
fined in this respect. East of San Diego the chaparral (mainly
Adenostoma fasciculatum) covers the gently sloping mesas, and is
solid on all exposures of the lower Cuyamaca Mountains (plate 11a).
The same is true, where environmental conditions are undisturbed,
in certain parts of the Los Angeles and San Bernardino Valleys and

16 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

on the foothills surrounding them. Farther north, in Coast Ranges
and Sierras, the chaparral is more and more restricted to south-facing
exposures (plate 10, a, b), the opposite slopes being occupied by
other plants, first by broad-sclerophyll trees, and still farther by
conifers and broad-leaved deciduous species as well. The same
changes are noted in ascending the higher mountains in any part of
the State.

LOCATION OF GREATEST IMPORTANCE OF THE TYPE IN THE

CLIMAX COMMUNITY.

This line of evidence is closely related to the last, since a plant
community which covers all slopes and exposures with fair uniformity
is likely to be the climax of that region. In another place (p. 72)
I will show that in certain regions the climax is dominantly of the
broad-sclerophyll type; that in others the broad-sclerophylls are
secondary and successional, and that there are intermediate areas
where the status of these plants is uncertain; and, finally, that the
region where the broad-sclerophyll type, and in particular the
chaparral, is most certainly climactic, is southern California.

CONTINUITY AND CONVERGENCE OF LINES OF DISPERSAL.
If we were to select an ideal center from which migrating species
might most quickly and easily reach all parts of the region where
broad-sclerophylls occur commonly, we would without fail fix upon
the vicinity of Ventura County. From this point easy migration
routes for all the species concerned, which would be mountain ranges
of moderate altitude, lead in various directions, and nowhere along
their courses are there barriers of any importance. From the north
two routes converge, the Coast Ranges, and the Sierras by way of
the Tehachapi Mountains. Southward are the various ranges
of southern and Lower California. Surely it is more than coincidence
that the evidence presented along other lines points to this same
region and the country immediately southeast of it.

I have endeavored in this chapter to present a picture of the range
of the California broad-sclerophyll vegetation-type, principally by
superposition of the ranges of the individual species; and to show,
through seven lines of evidence, that the center of distribution of
that type is in southern California west of the deserts, from the
Santa Ynez and Sierra Madre Ranges southward into northern
Lower California.

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 17

II. CLIMATIC RELATIONS.

Schimper (80) has shown clearly the constant relation between
sclerophyll dominance and climate. He says:

"The mild temperate districts with winter rain and prolonged summer drought
are the home of evergreen xerophilous wood}'- plants, which, owing to the stiffness of
their thick, leathery leaves, may be termed sclerophyllous woody plants.
Wherever original conditions have not been altered by man the sclerophyllous trees
and shrubs of districts with a moist winter always form dense and continuous wood-
land, which in most cases consists principally or exclusively of shrubs, but which
occasionally becomes true forest, although of low or middle height only."

As to the advantage of the evergreen habit, he states that the
vegetation is subject to short but frequent irregular periods of rest,
sometimes due to cold, sometimes to drought; that short periods
only afford simultaneously optimum conditions in temperature and
moisture; the absolute optima for these two factors being entirely
separate in time. It is therefore decidedly advantageous for the
plants to be prepared to do their assimilative and vegetative work
at all times. The regions which possess such a climate and support
such a vegetation are, according to Schimper, the Mediterranean
shores, the southwest extremity of Africa, southwestern and much
of southern Australia, central Chile, and California. More detailed
and localized works (4, 7, 8, 25, 78, 91) confirm Schimper's conclusions.

After such a thoroughly adequate, even though brief, treatment
by the pioneer author, it is necessary here merely to particularize
somewhat concerning the region under discussion.

The map (plate 2) is in part adapted from Reed and Kincer (77).
The iso-lines indicate the percentage of total precipitation occurring
in the half year April 1 to September 30. By comparing it with
the map (plate 1) a remarkable correspondence will be at once
evident between the region where the summer precipitation is less
than 20 per cent and the region where broad-sclerophyll species
are numerous. Moreover, the area of less than 10 per cent summer
precipitation corresponds closely with the center of distribution of
the sclerophylls, as described in the section devoted to that topic.
Seasonal distribution alone, however, is not sufficient to explain the
region of dominance of the broad-sclerophylls ; total precipitation is
only slightly less important. If the total is very high, or if atmos-
pheric conditions are such as materially to reduce evaporation during
the dry season, it may be possible for species that are less xerophytic
to control; if the total is very low, true desert species will replace
the broad-sclerophylls. I have therefore plotted, in a generalized
way, the total precipitation in the area with less than 20 per cent
summer rainfall. The iso-lines selected are of course arbitrary, but
they nevertheless mark out in a striking way the areas dominated
by the three great vegetation types. Where the total precipitation
is more than 30 inches, the vegetation is conifer forest of some type.

18 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

The particular reasons for its presence differ in different places.
In the Sierras and the isolated mountain areas of southern California
the dominating cause is high total precipitation, which furnishes a
sufficient amount of soil-moisture to tide the relatively mesophytic
vegetation over the unfavorable season. In the higher mountains
the slowly melting snows are of very great importance. In the
northern end of the State, in addition to a high total precipitation,
there is a decrease in the duration of the dry period. Along the
northwestern coast there is, in addition to the other factors, the
abundance of summer fog, which is the particular reason for the
redwood forest.

The area with 10 to 30 inches of rainfall is the region of broad-
sclerophyll dominance. The correspondence with the region deter-
mined upon through other lines of evidence is quite striking, even
the Sacramento Valley being included (see p. 81). Only the narrow
strip along the east slope of the Sierras, dominated by the vegetation
of the Great Basin, must be left out. In the north Coast Ranges
the broad-sclerophylls, both trees and chaparral, are of great impor-
tance, as plate 1 indicates. Observation shows, however, that the
conifers are competing with them on at least equal terms (see p. 73).
The climatic map, indicating conifer forest conditions for this region,
is therefore not deceptive.

Finally, those regions with less than 10 inches of rainfall correspond
accurately with the deserts — the western portions of the Colorado
and Mojave Deserts and the Owens Valley region — and in addition
the southern part of the San Joaquin Valley, which closely approaches
desert conditions.

The same points are brought out in another way by table 1.
The three great regions are seen by the summary to be thoroughly
distinct in total precipitation, the proportion being approximately:
desert 1; broad-sclerophyll 6; conifer forest 15.

The seasonal distribution in conifer forest and broad-sclerophyll
regions is much alike, the summer precipitation averaging well
below 20 per cent. In the desert the percentage is distinctly higher,
because a number of stations east of the 20 per cent line were included
in order to give a true picture of the California desert region as a
whole.

Certain temperature figures are added, though the data are incom-
plete and their interpretation unsatisfactory. The mean annual
temperature is of little use; the seasonal extremes, i. e., the means
of the hottest and coldest months (in nearly every case July and
January) and the mean maxima and minima of the same months
are more significant. The mean annual temperature of the broad-
sclerophyll region is only slightly below that of the desert. The
mean of the hottest month, on the contrary, is much lower, being

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 19

only slightly greater than that of the conifer forest. The striking
and significant fact is that the seasonal range of temperature is far
less in the broad-sclerophyll region than in either conifer forest or
desert, the minima especially being decidedly above the others.
This fact fits well with Schimper's elucidation of the evergreen
habit as a means of utilizing short periods of favorable conditions for
activity at all times of the year. The relatively narrow seasonal

Table 1. — Climate and vegetation.

Region.

Precipitation.

No.

sta-
tions.

Total
(inches).

P. ct.

May 1

to

Oct, 31. 1

Temperature (° C. ; No. of stations varies).

Mean
annual.

Mean

of
hottest
month.

Mean

of
coldest
month.

Mean

maximum

(hottest

month).

Mean

minimum

(coldest

month).

Conifer forest:

Redwood region

North Sierras

South Sierras

Transitional: North Coast

Ranges

Broad-sclerophyll :

South Sierra foothills. . . .

South Coast Ranges

Cuyamaca Mountains. . .
Grassland and coastal sage-
brush :

Sacramento Valley

Los Angeles-San Bernar-
dino Valley

San Joaquin Valley

Desert :

Owens Valley

Mojave Desert

Colorado Desert

Summary:

Conifer forest

Broad-sclerophyll

Desert

7

22

5

10

7
27
10

23

15
31

4
4

7

34
44
15

65.83
60.50
42.20

38.08

29.33
20.32
19.92

21.85

16.48
11.21

4.89
3.97
3.11

58.91

21.67

3.82

14

14.4

13.1

12.4

13.9

11
13.4

23.9

28.2
28.9

14.2
11.6

27

11.3
11.1

14

16.6
14.2
13.7

16.9

16.8
16.9

14.9
17.6

11.2
14.4
15.8

15.5

20.7

20.8

26.9
19.7
21.4

26.8

29.1

26.8

26.1
29.8

19.8
20.9
27.9

7.8
3.3

7.8

7.9

10.5
7.9

4.6
7.4

4.2
8.5
5.5

33.1

27.8
30.7

33
36

33.8
40.8

28.8
36.4

3.3
2.7
1.6

3.9
1.7

•4.0
1.6

1 The figures given here were worked out before the publication of the map (77). The difference in date
of beginning and ending of summer season does not appreciably affect the results.

temperature range is of course due to the coastal location of the
great mass of the broad-sclerophyll region, and it is of interest that
all the other regions dominated by this type of vegetation are coastal
too.

Summarizing briefly, we find that the broad-sclerophylls occur
abundantly in that part of western North America where the summer
precipitation is less than 20 per cent of the total, but that they
dominate only that portion of this area where the total precipitation
lies between 10 and 30 inches, and that the region of their dominance
is also characterized by very moderate temperature extremes.

20 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

III. THE COMMUNITIES.

The terminology of plant communities is just now in a state
of flux, and therefore it is too much to hope that the system employed
in this paper will be satisfactory to every reader. The best that
the author can do is to define his terms accurately and relate them
clearly to the uses of other writers. It is therefore necessary to
submit the following definitions:

The Formation. The fundamental unit of vegetation : a community
relatively homogeneous ecologically in the character of its dominants,
floristically in that its locally varying subdivisions are bound together
by the common dominance of one or more species or by the equiva-
lence of species ecologically near of kin, and developmentally in that
within a given climatic region it has a constant successional role.

The Association. A vegetation unit of lower rank than the
formation and contained within it; differing from other associations
within the same formation with respect to any or all of the bases of
the formation (ecological character, floristics, development), in
minor degree, but sufficiently to cause it to stand out as a distinct
entity.

These two units are analogous, respectively, to the taxonomic
units, species, and variety, and one may conveniently refer to the
range of a formation or association exactly as one speaks of that of a
species or variety. In naming the formations I would use descriptive
titles based upon ecological character, and for the associations I
would employ, when possible, the names of dominants.

The Consociation, an association with a single dominant, is fre-
quently a useful term. The word "community" is an indispensable
addition to the list, being used to designate any vegetation unit
without specifying its rank.

These terms must now be related to two recent systems, that of
Clements (21) and that of Nichols (69). Clements applies the
term formation to the climax community only; I would extend it to
successional communities as well. Clements distinguishes between
the climax units association and consociation and the successional
associes and consocies; my use of association and consociation includes
both types of communities. The formation of the present paper
agrees in substance with the association-type of Nichols; in the use
of the terms association and consociation we are in essential agreement.

THE BROAD-SCLEROPHYLL COMMUNITIES.

There are two Californian formations in which the broad-sclero-
phylls are the dominating element — the broad-sclerophyll forest
formation and the chaparral formation. Each is in part climax,
in part successional. Further, there is a broad-sclerophyll element
of minor importance in the redwood forest, making a rather large

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 21

part of its undergrowth. The grounds upon which the formations
have been distinguished, and their range, composition, and structure,
will be given here. Discussion of their relations to climate and to
other communities is included in other chapters.

The broad-sclerophyll forest formation is dominated by trees,
mainly sclerophyllous evergreens, but including a number of decid-
uous species (30.8 per cent). It is typically climax, but in this phase
its extent is limited. It is successional where its range overlaps the
ranges of the conifer formations, and postclimax in its overlap with
the climax chaparral.

The chaparral formation is made up of shrubs, the great majority
being sclerophyllous evergreens. Its climax and successional phases
are both of great importance. The latter is related developmentally
to the conifer climaxes and is almost totally distinct floristically from
the climax phase.

THE BROAD-SCLEROPHYLL FOREST FORMATION.

This formation ranges from southern Oregon southward through
the coast mountains and Sierra foothills into Lower California,
reaching its limit probably in the region of Mount San Pedro Martir.
Nowhere, so far as I am aware, does it dominate the country as a
conifer forest, for instance, commonly does. It occurs rather in
discontinuous patches, which may, however, be of considerable
extent. These alternate in the main with patches of chaparral of
the type which I have designated as climax. Northward the forest
is the more important of the two, especially in the Coast Ranges,
while southward the chaparral becomes more and more preponderant.
In the north there is overlap also with the ranges of the Sequoia
sempervirens and the Pseudotsuga associations of the Pacific conifer
formation, and in the Sierras with the formations of the conifer
forest region.

The number of dominant species is not large. In the following
list a single asterisk indicates importance also in the conifer forest
chaparral; and two asterisks, in both that and the climax chaparral.

Sclerophylls.

Myrica californica. Quercus chrysolepis. Pasania densiflora.*

Castanopsis chrysophylla.* engelmanni. Umbellularia californica.

Quercus agrifolia. wislizeni.** Arbutus menziesii.

Deciduous.
Quercus kelloggii.* Acer macrophyllum Aesculus californica

lobata

Several associations occur, easily recognized because of their
relative constancy and wide distribution. Many individual localities
would fit into none of them, and therefore in a minute study numerous
transitional units might be described.

22 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

Pasania-Quercus- Arbutus Association. — This community is charac-
teristic of the lower altitudes of the North Coast Ranges, a region
very complex and rather difficult to understand vegetationally,
since two or more types which are climactic nearby meet here and
overlap, finding conditions that are reasonably favorable to all.
The redwoods thoroughly dominate the coast. East of them
Pseudotsuga mucronata is the commanding species, but shares its
rule with the broad-sclerophyll association about to be described.
Chaparral and grassland communities also occur, but these are
plainly successional. The Pasania-Quercus- Arbutus association is
itself somewhat of a transitional unit between broad-sclerophyll
and conifer types, for it rarely occurs without at least a sprinkling of
conifers, especially Pseudotsuga, and its principal species occur
commonly as an understory of the Pseudotsuga and Sequoia forests.

The most important species of the formation are Pasania densiflora,
Quercus chrysolepis, and Arbutus menziesii, and these attain great
size. Other tree species occurring more or less commonly are
Quercus kelloggii, Castanopsis chrysophylla, Umbellularia californica,
Acer macrophyllum, JZsculus californica, and Cornus nuttallii. The
association ordinarily possesses two layer societies — one of shrubs,
including Corylus rostrata californica, Vaccinium ovatum, and Gaul-
theria shallon, and one of herbs and ground-shrubs, a mixture of
typically oak-forest species and those commonly associated with the
redwood and douglas fir.

The transitional phases of the association will be made evident
by description of two areas, one in the interior of the Coast Ranges,
where Pseudotsuga is the competing tree, the other on the edge of
the coastal redwood region.

The first locality is in Trinity County, on the north-facing slope
of the Mad-Trinity Divide. The dominant tree, Pseudotsuga, grows
here magnificently, many specimens attaining a diameter of 2
meters. Abies concolor and Pinus lambertiana also occur. Beneath
the conifers there is an understory of broadleaf trees, nearly all
sclerophylls. Castanopsis chrysophylla is the most abundant, and
Pasania, Arbutus, and Acer macrophyllum also occur. This assem-
blage might here be termed a layer society. Corylus rostrata cali-
fornica, Cornus nuttallii, and Ceanothus integerrimus form a second
stratum, and a third is composed of herbs and ground shrubs:
Vancouveria sp., Polystichum muniturn, Berberis sp., Gaultheria
shallon.

The other is the valley of the South Fork of the Eel River, in
Humboldt County. In the vicinity of Garbersville and for several
kilometers north of it (downstream), the north slopes and ravines
are forested with Pasania densiflora, Quercus kelloggii, Q. chrysolepis,
Q. garryana, Arbutus menziesii, Umbellularia californica, Msculus

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 23

californica, Acer macrophyllum, and Pseudotsuga, making a rather
typical specimen of the Pasania-Qucrcus- Arbutus association. The
first redwoods appear in groups of large trees on the valley bottom,
with scattered individuals on north slopes (plate 8a). This con-
tinues for 15 kilometers or more, then for several kilometers there is an
almost pure forest of Sequoia on north slopes with the Pasania-
Quercus- Arbutus association on south exposures. Finally the forest
becomes nearly pure Sequoia on all slopes, the broad-sclerophylls
gradually disappearing as a distinct community, though remaining
to some extent as a layer society, particularly Arbutus and Pasania.

Quercus agrifolia- Arbutus Association. — This unit is coastal, occur-
ring from the northern limit of Quercus agrifolia in Mendocino County
to the southern limit of Arbutus in Los Angeles County. It is thus
characteristic of the outer central Coast Ranges, where it is the
dominant cover on north-facing slopes. It is particularly well
developed in the San Francisco Bay region and southward to the
Santa Lucia Mountains. The character tree is Quercus agrifolia. 1
Arbutus menziesii is next in importance, but varies greatly in
abundance in different localities. ^Esculus californica, a deciduous
species, is usually prominent, and Umbellularia californica is equally
so. Acer macrophyllum is frequently important in the more meso-
phytic localities. In areas that are transitional with the Sequoia
association, Pasania, Quercus chrysolepis, Q. kelloggii, and Sequoia
itself occur. Since a typical area of this association is described in
another part of this paper (p. 38), it will be unnecessary to go further
into details here. Station 7 at Jasper Ridge is representative in
every respect (plates 14a, 8b).

Quercus agrifolia Consociation. — South of the southern limit of
Arbutus (Los Angeles County) the community is continued as a
consociation dominated by Quercus agrifolia. This is rather promi-
nent in the lower altitudes of the west slope of the Cuyamaca Moun-
tains.

Umbellularia Consociation. — Umbellularia occurs scattered through
the Quercus agrifolia-Arbutus association, and in others as well, but
it also forms pure growths, especially in the central Coast Ranges,
occupying moist ravines and canyon bottoms. These groups of
Umbellularia stand out strikingly above the other trees, being con-
spicuous by reason of their light green color and conifer-like form.
The shade is very dense and undergrowth almost lacking. Umbel-
lularia itself, however, is able to germinate successfully under such
conditions.

Quercus agrifolia-lobata Association (plate 9a). — This is charac-
teristic of the broad valleys and gentle footslopes of the central

1 In some places, especially in the inner Coast Ranges, it is replaced by Q. wislizeni, and thus
another association might be distinguished.

24 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

Coast Ranges, being locally of considerable importance in the San
Francisco Bay region. The dominant species are Quercus agrifolia
and Q. lobata, the latter being deciduous. The trees as a rule stand
far apart, producing a park-like landscape, and it is in such places
that the largest specimens of both species occur. One tree of
Q. agrifolia near Palo Alto is 2.1 meters in diameter breast-high.
A specimen of Q. lobata west of Clear Lake is of the same diameter,
with a spread of branches of 47 meters. Much larger trees of the
latter have been reported. Other species are of occasional occurrence.
In the Palo Alto region large specimens of Umbellularia, Arbutus
and Prunus ilicifolia make a small part of this association. Because
of the wide scattering of the trees, the ground between is in most
places under cultivation. Near Palo Alto, however, there are a
few localities which retain their original vegetation, because they
have long been included in certain large estates. In such areas one
finds the two oaks of all sizes from seedlings to large mature trees.
Most of the young ones occur in indefinite groups in the opener places,
while the mature specimens completely dominate the ground beneath
them. Three layer societies occur. The first, of tall shrubs, includes
Rhamnus californica, Heteromeles arbuiifolia, Sambucus glauca,
and Rhus diversiloba. The low-shrub society includes Rubus viti-
folius, Symphoricarpos racemosus, and Solatium umbelliferum. Mic-
romeria chamissonis is dominant in the ground layer society. Further
details concerning this very interesting association will be given in a
future paper upon the communities and successions of the Palo
Alto region.

Quercus chrysolepis-kelloggii Association. — The associations so
far described are distinctly of low altitudes. The present one belongs
to the higher Coast Ranges and southern California mountains and
to the middle altitudes of the Sierras. It is preeminently a north-
slope forest, but localities are common enough where it occurs on
other exposures as well, seeming like a true climax. The most
important tree species is the broad-sclerophyll Quercus chrysolepis;
Q. kelloggii, deciduous, is often a close second. Others are Arbutus,
Umbellularia, Acer macrophyllum, Pasania, ALsculus. Since the
association has so great a range, the subordinate vegetation varies
greatly. It shows broad transition areas with neighboring associa-
tions. Its close relation to the Pasania-Quercus- Arbutus association
is at once evident, and it is not strange, therefore, to find areas that
can not be placed with certainty in either. Again, just as the
Pasania-Quercus- Arbutus association passes into the Sequoia forest
as an understory, so also does the Quercus chrysolepis-kelloggii
association into the pine forests of the high Coast Ranges and the
Sierras. In fact, in the Sierras it is commoner to find the community
as an understory beneath Pinus po?iderosa and P. lambertiana than

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 25

as a dominating type. In the mountains of southern California
the group forms a similar understory beneath Pseudotsuga macro-
carpa. Upon the xerophytic side there is transition to the chaparral.
Such areas have so individual a stamp that I have been accustomed
to refer to them as "dwarf forest." An excellent example is found
upon the north slope of Mount Tamalpais (Marin County), near
the summit. Quercas chrysolepis and Q. wislizeni, growing in dense
thickets 3 to 5 meters in height, are dominant, an occasional full-
sized tree of Q. chrysolepis rising above the general level. With
them grow other species: Quercus agrifolia, Pasania, Arbutus,
Umbellularia, Torreya californica, and the chaparral shrubs Arcto-
staphylos tomentosa [A. glandulosa Eastwood], Ceanothus sorediatus,
and Castanopsis chrysophylla minor.

Quercus chrysolepis Consociation (plate 9b). — In the middle
altitudes of the Sierras, dominated by the pines and Pseudotsuga,
Quercus chrysolepis growing almost pure has a distinct successional
role. Upon the great talus accumulations at the bases of the Yo-
semite cliffs certain chaparral shrubs are the pioneers. These are
followed by a dense, pure growth of Quercus chrysolepis which seems
to persist for a long time, as the live-oak forest is the most conspicuous
feature of such areas. The talus piles that are manifestly oldest,
with much accumulation of humus, support a mixture of the oak
and Pseudo suga.

A few concluding remarks in summary will gather together the
main points in the discussion of the broad-sclerophyll forest formation.
It is plain that the group as a whole is the fundamental unit, the
minor divisions being closely tied together by a number of binding
species. The transition zones between associations of the formation
and with other formations are broad, so that accurate delimitation
is difficult. The broad-sclerophyll communities, wherever they
adjoin the conifer forest communities, pass into them as layer societies.
There is a very close habitat relation between the broad-sclerophyll
forest and the climax chaparral, in that in the main they overspread
the same range, occupying areas of comparatively slight physical
differences. The question of climax, therefore, whether one or the
other or both, is difficult. My conclusions will be given in a later
section.

THE CHAPARRAL FORMATION.

It was pointed out at the beginning of this chapter that the
chaparral formation, homogeneous in the ecological character of its
dominants, at least so far as anatomical structure is concerned,
embraces two associations which are very distinct floristically and
developmentally. Separate treatment of these will be the best
method of presentation.

26 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

The Climax Chaparral Association.

The climax chaparral is the dominant community over the whole
of the southern Coast Ranges and the mountains of southern Cali-
fornia and northern Lower California. Only the highest summits,
controlled by conifers, and the more mesophytic north slopes,
inhabited by broad-sclerophyll forest, must be excepted. North-
ward in the north Coast Ranges the chaparral shares its control
more and more with the broad-sclerophyll trees, and opposite the
northern end of the Sacramento Valley it disappears entirely as a
dominating community. In the southern Sierras it is of great
importance, occupying a wide belt in the foothills. In the northern
Sierras its continuity is broken, and this is not strange, since the
conifers of the montane forest here reach the valley floor. The
present range of the climax chaparral is indicated in a very general
way by the range of its most important species, Adenostoma fas-
ciculatum (plate 3). In addition, I believe that there are certain
extensive areas now inhabited by grasses and by half-shrubs that
climatically and potentially are chaparral regions (p. 76).

Since the climax chaparral is by far the most widely extended
and diversified of the broad-sclerophyll communities, it is natural
that the present list of species should be the longest. The following
are all evergreen, except that Quercus dumosa is barely so. One
asterisk (*) indicates that the species is also of importance in the
conifer forest chaparral; two asterisks (*) that it is important in that
and also in the broad-sclerophyll forest.

Castanopsis chrysophylla minor.*
Quercus chrysolepis.**

dumosa.

durata.

wislizeni frutescens.**
Dendromecon rigidum.
Heteromeles arbutifolia.
Cercocarpus betula?folius.
Adenostoma fasciculatum.
sparsifolium.
Prunus ilicifolia.
Xylothermia montana.
Cneoridium dumosum.
Rhus integrifolia.

Rhus laurina.

ovata.
Rhamnus californica.*

crocea.
Ceanothus crassifolius.

cuneatus.

dentatus.

divaricatus.

hirsutus.

megacarpus.

papillosus.

rigidus.

sorediatus.

verrucosus.

Garrya elliptica.
Comarostaphylis diversifolia.
Xylococcus bicolor.
Arctostaphylos andersonii.

glauca.

hookeri.

manzanita.

montana.

pumila.

stanfordiana.

tomentosa.

vestita.
Eriodictyon californicum.

With so large a list of species there is naturally great diversity in
the composition of the association. Anyone given to splitting of
hairs would easily separate many communities of lower rank. This
is in part due to slight habitat differences, but also in an important
degree to the great number of species with restricted range and to
the frequent occurrence of fires, which result in multitudinous com-
binations of species, depending upon which are able to survive or to
repopulate the area burned. It is easy to recognize, however,
throughout the length and breadth of the region, one striking and

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 27

characteristic consociation, for Adenostoma fasciculatum covers
many hundreds of square miles in practically pure dominance.
The range of the Adenostoma consociation is indicated on the map
(plate 3). Other species, too, completely control certain areas,
but it is far commoner for these to mingle with each other and with
Adenostoma in an endless number of combinations and proportions.
In 87 listed localities in all parts of the State, the following occur-
rences of important and widespread species are noted:

Adenostoma fasciculatum 75

Arctostaphylos (all species) 50

Heteromeles arbutifolia 26

Ceanothus cuneatus 25

Quercus dumosa and Q. durata 23

Cercocarpus betulsefolius 19

Quercus wislizeni frutescens 11

Rhamnus calif ornica 10

Quercus chrysolepis 10

Such being the condition, it avails little to attempt to distinguish
minor units within the association. It is more reasonable to express
the differences by noting the dominance of one or more species in
particular cases. One fact, however, must be brought forward.
The genus Arctostaphylos gives its stamp to certain localities in a
very characteristic way. No one species is dominant throughout.
Arctostaphylos tomentosa is by far the most important, ranging over
the whole region. A. glauca is abundant in the southern half of the
State and A. manzanita in the northern, and several others are
prominent locally. This phase nearly everywhere accompanies
the Adenostoma consociation, occupying the less xerophytic north-
facing slopes where these are not sufficiently moist to permit the
forest to exist, and at higher altitudes replacing the Adenostoma
consociation on the south slopes, the north exposures being forested.
The combination is well shown at Jasper Ridge, described in the
next chapter.

The Conifer Forest Chaparral Association.

The range of this community is nearly coextensive with that of
the montane conifer forest, spreading to some extent into the region
of the subalpine conifer forest. Its home is, therefore, in the middle
altitudes of the Sierras, with extensive colonies throughout the
higher mountains of northern California and Oregon, the north
Coast Ranges, and the mountains of southern California. It reaches
its best development where the forest which controls it is best
developed, and therefore its true center is in the northern Sierras.
Accurate altitudinal data are unavailable, and the patchy distribution
makes close determination impossible, but the lower limit may be
placed roughly at the lower limit of Pinus ponderosa, which is very
near to the valley-level at the northern end of the Sierras and rises
gradually southward. Naturally there is extensive overlap with
the range of the climax chaparral which dominates the foothill
region in imperfect manner. The upper limit is equally indefinite.

28 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

Above the ranges of the dominant trees of the Montane Forest the
chaparral species become gradually fewer. At least one (Arctosta-
phylos nevadensis) reaches timberline in the Yosemite region (39).

The following list includes those species which are of importance
in the conifer forest chaparral in various parts of its range. Many
others, of course, occur occasionally, especially where the range of
the formation overlaps that of another. Species marked with a
single asterisk (*) are of importance also in the broad-sclerophyll
forest, those with two asterisks (**) in the climax chaparral, and
those with three in both.

Sclerophylls.

Castanopsis chrysophylla minor.* Pasania densiflova echinoides.* Arctostaphylos mariposa.

sempervirens. Rhamnus californica.** 1 nevadensis.

Quercus ehrysolepis.*** Ceanothus cordulatus. patula.

sadleriana. velutinus. pungens.

vaccinifolia. Garrya fremontii. viscida.

wislizeni.*** Arctostaphylos drupacea.

Deciduous.

Corylus rostrata californica. Amelanchier alnifolia. Cercis occidentalis.

Quercus breweri. Prunus demissa. Acer glabrum.

garryana. emarginata. Ceanothus integerrimus.

kelloggii.* subcordata. sanguineus.

The conifer forest chaparral association is at least as variable as
the climax chaparral. This is due in part to great variation in habitat
and also in an important degree to what may be termed accidental
causes, for since the community owes its existence in large part to
the frequent destruction of the forest by fire, the population of a
given area must depend largely upon what species are producing
seed in its immediate vicinity. The association has been carefully
studied by the U. S. Forest Service, since its occurrence within the
belt of merchantable timber gives it great economic importance.
A number of reports have been published (see introduction) and
many more are still in manuscript, and to these I have had access
through the kindness of the officials of District 5 in San Francisco.

Since the work of the Forest Service upon this community greatly
outweighs mine, and since more of it is soon to be published, I will
not attempt detailed subdivision. A few general features are worth
pointing out, however. In the lower and drier portion of the yellow-
pine belt of the Sierras one combination is of great importance.
This is made up of Arctostaphylos viscida and Ceanothus integerrimus,
the former being especially conspicuous because of its beautiful gray
foliage. It is particularly widespread in the regions surrounding
the old placer diggings, where the timber was cat and burned
off more than half a century ago. In the upper and moister por-
tion of the pine belt and the lower part of the Subalpine Zone
the areas of chaparral show greater floristic diversity, and a great

1 Principally in its two vaiieties, R. c. tomentella and R. c. rubra. The latter is deciduous.

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 29

number of consociations and mixtures might be distinguished.
In the higher northern Sierras half a dozen species are of paramount
importance. These are Castanopsis sempervirens, Quercus vac-
cinifolia, Prunus emarginata, Ceanothus cordulatus, C. velutinus,
Arctostaphylos nevadensis, and A. patula. They occur in all possible
combinations, so that a classification for one region may not fit
another at all. We should also mention the thickets of deciduous
oaks, especially Quercus kelloggii and Q. garryana, which are so
extensive in the mountains of Trinity County and northward.
These species, ordinarily trees, here grow in dense masses like the
chaparral, forming a connecting link ecologically with the deciduous
oak thickets of the Rocky Mountain foothills.

The following outline of the broad-sclerophyll communities will
serve as summary to this chapter, all being climax except those
specially noted:

A. Broad-sclerophyll forest formation.

1. Pasania-Quercus-Arbutus association.

2. Quercus agrifolia-Arbutus association.

2a. Quercus agrifolia consociation.
26. Umbellularia consociation.

3. Quercus agrifolia-lobata association.

4. Quercus chrysolepis-kelloggii association.

4a. Quercus chrysolepis consociation (successional) .

B. Chaparral formation.

1. Climax chaparral association.

la. Adenostoma and other consociat'ons and various mixtures.

2. Conifer forest chaparral association (successional).

2a. Numerous vague minor units.

30 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

IV. VEGETATION AND HABITAT.

THE LOCALITY FOR INTENSIVE STUDY.
THE PALO ALTO REGION.

For intensive study of vegetation and habitat it was necessary
to find a spot as nearly representative as possible, affording oppor-
tunity for comparison of related communities and favorably situated
with respect to a base of operations. The vicinity of Palo Alto
satisfied these requirements, the botanical laboratory of Stanford
University furnishing facilities for indoor work.

Palo Alto is situated in the central Coast Range region, near the
southern extremity of San Francisco Bay, from which it is but
3 km. distant. The lowland upon which it lies extends along the
southwest shore and is commonly spoken of as a northwestward
extension of the Santa Clara Valley. Southwestward from the town
rise the Santa Cruz Mountains, the crest, ranging from 512 to 850
meters in height, being 11 km. distant. The northeast face of the
range is a fault scarp, and therefore abrupt. There is, nevertheless,
a zone of foothills between the main mountain front and the valley,
in the vicinity of Palo Alto 5 km. broad and attaining a height
of 246 meters. The altitude of Palo Alto itself is but 17 meters.

The topographic diversity just outlined, together with the influence
of the ocean, which is distant from Palo Alto 25 km. over the Santa
Cruz Range, produces very great climatic differences within short
distances. Thus, the annual precipitation at Palo Alto is 66 cm.,
while on the mountain summits, only 15 km. away, it is 139 cm.
(23, p. 186). At the ocean shore the precipitation is again light,
75 cm. being recorded at one station. The summer fogs, which
regularly cover the mountains but rarely invade the valley, are also
of profound climatic importance.

In correlation with topographic and climatic diversity we find
equal diversity in the vegetation. The prevailing type in the fog-
bathed mountains is the redwood forest ; the well-watered mountains
without fog support a vegetation made up of evergreen oaks, madrono,
and Douglas fir; the foothills and drier mountain sides are clothed
with chaparral except where local conditions, such as steep north
slopes, permit the development of oak forest; the gentle slope from
foothills to bay was originally covered by chaparral, open oak forest,
and salt marsh, all of which have been more or less disturbed by
cultural operations. For an account of the redwoods and their
relation to rainfall and fog, the reader is referred to a former paper (23).

JASPER RIDGE.

The locality chosen for quadrat and instrumental study was one
of the higher foothills, 7 km. southwest of Palo Alto, known locally
as Jasper Ridge. It is a hilly mass, approximately 10 km. northwest

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 31

and southeast, by half as much in the other dimension. It is nearly
surrounded by the valleys of San Francisquito, Los Trancos, and
Corte de Madera Creeks, the floors of which range in altitude from
60 to 120 meters. The average height of the crest of the ridge is
approximately 180 meters, and the highest point, near the south-
eastern end, 246 meters. The mass here has a rather broad, rolling
summit, but the slopes are considerably dissected by ravines. North-
ward and northeastward from Jasper Ridge the hills are lower, but
southwestward, across the valley of Corte de Madera Creek, the
main front of the Santa Cruz Mountains rises to an altitude of 600
meters.

The surface rock of the main part of the ridge is probably the
Chico sandstone of Upper Cretaceous age (14). There are two
areas of the Franciscan formation of considerable extent and one
outcrop of serpentine, but the area studied is entirely within the
limits of the presumptive Chico. This rock weathers into a uniform
coarse yellow sand. In most places the layer of residual soil is
very thin, but on a few level spots there is a greater depth. In
excavations in such places the sand is seen to merge gradually into
the undecomposed rock, which is found practically intact at a maxi-
mum depth of a meter or a little more. On steep hillsides the soil
is irregular in depth, with rock outcrops alternating with sand
pockets. The greatest accumulations are naturally at the foot of
the steepest slopes, where the sand contains many angular fragments
of considerable size. Of the ten stations studied, eight possess a
uniform soil of the type just described. The other two show local
differences which are of sufficient importance to affect considerably
the vegetation growing on them.

At first thought it would seem absurd to devote a paragraph to
the climate of so limited an area as Jasper Ridge, after the climate
of the region as a whole has been discussed. In California, however,
such is not the case. One can not assume that the climates of two
places are alike because they are but a few kilometers apart. Witness
the great difference in rainfall already noted between Palo Alto
and the summit of the Santa Cruz Range. In the matter of rainfall,
Jasper Ridge lies between the above stations, as it does in position.
The precipitation for a number of stations in the Santa Clara Valley,
foothills, and Santa Cruz Mountains in the season 1913-14 (a very
wet year), was as follows (see 23, p. 185):

cm.

,, „ ( San Jose 47 . 83

Valley: | Palo AltQ 62 15

Foothills: Jasper Ridge 103 .71

, , . . ( King's Mountain 169 . 25

Mountains :{ Ben Lomond ; 197 66

Assuming that the proportion between the stations will hold
roughly constant for a succession of years, the normal rainfall at

32 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

Jasper Ridge, calculated from those of San
Jose and Ben Lomond Weather Bureau Sta-
tions, is about 80 cm.

It seems probable that Jasper Ridge has a
smaller total of fog than either the valley or the
mountains. The true ocean fogs rarely reach
it and the "bay" or "tule fogs," which are
frequent during the forenoon in the lowland,
do not ordinarily rise so high. The average
midday temperature is probably higher than
in the valley because of the proximity of the
bay to the latter, and also higher than in the
mountains because of altitude and fog. Tem-
perature and fog affect vegetation through their
influence upon the evaporating power of the
air, and it is quite certain that the evapora-
tion-rate is much higher upon Jasper Ridge
than in the mountains, and almost as certain
that it is higher than in the valley, because of
distance from the bay, absence of fog, and prob-
ably higher mid-day temperature. We there-
fore conclude that Jasper Ridge is far more
xerophytic in climate than the mountains and
perhaps somewhat less so than the valley. It
may be, however, that the greater rainfall is
neutralized by the higher evaporation-rate.
The evidence of the vegetation is that there is
little climatic difference between the ridge and
the valley.

THE VEGETATION.

The vegetation of Jasper Ridge was orig-
inally mainly chaparral. Large portions have
been cleared of the bushy growth, quite cer-
tainly within a century, and the cleared areas
are now expanses of wild oats (Avena fatua
and A. barbata), with scattered oaks, especially
Quercus douglasii, which is peculiarly charac-
teristic of such secondary areas. Quercus agri-
folia and Q. kelloggii are also frequent, and
there is a considerable scattering of shrubs of
the species that frequent such cleared areas:
Ceanothus cuneatus,Heteromeles arbutifolia,Bac-
charis pilularis, and especially Rhus diversiloba.

Fig. 1. — Profile of a part of Jasper Ridge, drawn approximately
to scale, showing topography along trail, distribution
of plant communities, and location of stations.

>*

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

33

The original vegetation still remaining is comprised in several
large patches, the two principal ones being at the northwest and
southeast extremities of the ridge. It is true that these are mere
remnants, but they are of sufficient size to show absolutely natural
conditions, except along the edges. The studies were made in the
area at the southeast extremity, which covers approximately 2.5
sq. km. and includes the highest point of the ridge. It possesses
rather bold relief, including several summits and ravines, one of
which is decidedly abrupt. The extreme range of altitude is from
184 meters at the bottom of the deepest ravine to 246 meters at

A'

a'
A 9 A

A'

A'

A 3

A** A

A'

a '5

A'

A 5

A'

A

A 1

A , A

A' a-5

A A

A'

3

A' A

a' a
, A A 3
A 2

A'

A

A'r

.10

a' A'

A 3

A^

A „2

A

A

A

A A n

/

A

t

A

A'

A* A

A' 3

A

A 3A

12

A 3

A 3

A 3
A" ,

Ar

4

A'

A 5

. C

A A'

A v

A 2

A?

A 4 r

A 5

A'

A 3

A'

a' ,

Ar

k4

a'

A 3

A A', A 2

A 7
A'
A'

A 3

12

Ar

5 A

A A 3
A 2

A 2

A 6

A 3

A 6

A'

A'

A'r A '

Fig. 2. — Quadrat at station 1, 5 meters square Note abundance
of individuals of Adenostoma, which, however, are
small and do not control the area. For symbols in
this and succeeding quadrats see footnote, p. 34.

the highest hilltop. The vegetation of the area is mainly chaparral.
The Adenostoma consociation covers the south-facing slopes and the
summits and is by far the most extensive type; Arctostaphylos is
dominant upon the gentler north-facing slopes; and on the north-
facing slope of the deepest ravine is a typical area of the Quercus
agrifolia- Arbutus association of the broad-sclerophyll forest. It is
thus an ideal place in which to study the various communities of
broad-sclerophyll vegetation and their relations to habitat and to
each other.

For quadrat and instrumental study a series of ten stations was
determined upon, lying on an approximately north-south line

34 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

which crossed all slopes and vegetation types, giving two or more
examples of each. A trail was cut through the brush connecting
these with each other and with a nearby road. The distance between
the extreme stations was approximately 800 meters.

Stations 7 and 8 represent the forest and the remaining eight the
chaparral. Of the latter, stations 1, 4, 5, 9, upon south-facing slopes,
and 2 and 10, on nearly level summits, were in areas of Adenostoma
dominance, and stations 3 and 6, upon north-facing slopes, in Arc-
tostaphylos dominance. Quadrats were charted in six of these.
A number of statistical quadrats confirm the results.

Station 1.

This has the poorest vegetation of any. A quadrat 5 meters
square is represented in figure 2. 1 The species within it are given
below in order of abundance, with the number of plants of each
and the number of individual stems. Exact statistical study of
chaparral vegetation is difficult, because of the common habit of
growing in clumps and the frequent impossibility of readily deter-
mining the limits of a single group. The figures given are therefore
not absolutely accurate.

Clumps. Stems.

Adenostoma fasciculatum 81 219

Arctostaphylos tomentosa 9 60

Total 90 279

Heteromeles arbutifolia is occasional nearby. The bushes are all
small, ranging from 3 to 12 dm. in height, most of them nearer the
former figure. In spite of the large number of individual plants,
the ground is poorly controlled, large areas being entirely unshaded.
Herbaceous growth is practically absent, the quadrat showing not a
single plant at the time when it was plotted (August 25). There is
no humus superficially visible, and very scanty litter.

Station 2.

This station is on the highest point of Jasper Ridge and there-
fore most thoroughly exposed to atmospheric agencies. Moreover,
with station 1, it differs in soil character from the other eight, as
will be later shown. The vegetation is of better appearance than
that of station 1, but is still low and scattered, so that it is possible to
walk between the bushes in many places. No chart quadrat was

1 The symbols used in this and the following quadrats are given here. Exponents, except
in figure 6, indicate number of stems in a clump.

A. Adenostoma fasciculatum. B. Baccharis pilularis. Qa. Quercus agrifolia.

Ab. Arbutus menziesii. C. Ceanothus sorediatus. Qd. Quercus durata.

Ae. ^Esculus californica. D. Diplacus glutinosus. Qw. Quercus wislizeni.

Ar. Arctostaphylos tomentosa. H. Heteromeles arbutifolia. R. Rhus diversiloba.

As. Aster radulinus. Ho. Holodiscus discolor.

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

35

made, but a statistical quadrat of 5 meters square gave the following
results :

Clumps.

Adenostoma 98

Arctostaphylos 4

Total 102

Quercus durata is rather frequent in the vicinity and Heteromeles
arbutifolia and Quercus wislizeni also occur. Undergrowth is hardly
more noticeable than in station 1 . There is practically no humus and
very scanty litter.

Station 3.

A short distance north of station 2 a moderate slope begins, which
extends downward for a distance of 75 meters to the bottom of a
shallow ravine. From the summit for a little distance downward,

X-Ar

Qd

Ar

Qd

Ar

8 i

H H

Ar

Ar <
A

Ar

Qd

Ar

A Ar

Ar

Ar

Ar

Ar

Ar

Ar

Ar

Ar

Qd

Ar

Fig. 3. — Quadrat at station 3, 5 meters square. Note fewness of
individuals, mostly Arctostaphylos, which nevertheless
thoroughly control the area.

Adenostoma is predominant, but a gradual change is noted, until
the lower half is seen to be dominated by Arctostaphylos and Quercus
durata, with Adenostoma a decided minority. The size of the shrubs
also increases gradually, until near the bottom of the slope they
average 2 meters in height. The density is such that the ground is
entirely dominated, and the only way to travel is on one's hands
and knees below the zone of tightly intertangled branches. The

36 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

enormous woody masses which frequently form the bases of the
Arctostaphylos and Adenostoma clumps are very prominent here,
occupying a surprising amount of the ground area. The quadrat
shown in figure 3 gives the following summary:

Clumps. Stems.

Adenostoma 7 23

Arctostaphylos 18 44

Quercus durata 4 8

Heteromeles 3 11

Total 32 86

An occasional plant of Diplacus glulinosus, a half-shrub, is found
here, badly off for light. Depauperate specimens of Symphoricarpos
racemosus are rather frequent. Herbaceous growth occurs, but not
in great amount. The most abundant is Aster radulinus, which
produces nothing but basal leaves under the solid chaparral cover.
The only other is the fern Gymnogramme triangularis. The woody
bases of the shrubs are covered with mosses and small Cladonias.
Humus is in fair amount and litter is abundant. Where the trail
crosses the ravine bottom a single young specimen of Quercus agri-
folia suggests a tendency toward mesophytic forest conditions.

Station 4.

Crossing the ravine, the vegetation is seen to change, with no
transition zone, to Adenostoma dominance. The striking contrast
is manifestly related to the sharply angular change of slope at the
ravine bottom. Station 4 is exactly opposite station 3 and at an
equal height. The bushes, averaging 1.2 to 1.5 meters, are decidedly
better in appearance than those of stations 1 and 2, and the ground
is controlled by them to a much greater degree, though not completely.
The quadrat (fig. 4) contains the following:

Clumps. Stems.

Adenostoma 24 133

Arctostaphylos 5 42

Quercus durata 3 4

Total 32 179

The undergrowth is exceedingly sparse, including a few individuals
of Aster radulinus and Gymnogramme triangularis. Humus is
scanty, but there is considerable litter beneath the shrubs. A plant
of Quercus wislizeni close to the quadrat has three trunks 2.5 meters
high, standing well above everything else.

Station 5.

Going northward up the gentle slope, we pass through vegetation
like that of station 4, but becoming lower and less dense. There
are frequent areas where one may walk between the bushes, many of
which are of low stature. Adenostoma is everywhere dominant,
and Arctostaphylos, Heteromeles, Quercus durata, and Q. wislizeni are
also present. Undergrowth is practically absent, and humus and
litter are scanty.

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 37

Station 6.

Passing over the second summit, we find Adenostoma still dominant

and greatly increasing in size and controlling power. Descending

the steep north slope a short distance, we encounter a gradual change

to conditions like station 3. Here station 6 was established. The

Qd

Qd

2

A

A^ A<
A J

A
A A

31

Qd

;

Ar

,*

A"

A

v_

^

2

Ar

A J

a'

;

A

13

A £

Ar

A

A

4

A

A 4

IB

, A
A

Qw

7

A
a
A

A 7

a' 5

6

A 4 A '

A

Fig. 4. — Quadrat at station 4, 5 meters square. Note relative
fewness of individuals of Adenostoyna, which control
the area with fair completeness. The dotted line
indicates the area dominated by the large clump of
Arctostraphylos (Ar 34 ).

shrubs are taller than in any previous station, averaging 1.8 meters,
and thoroughly control the ground. The summary of a quadrat
(fig. 5) follows:

Clumps. Stems.

Adenostoma 2 8

Arctostaphylos 20 55

Quercus durata 2 4

Heteromeles 1 6

Total 25 73

The bases of the stems have the large woody masses noted in
station 3, and the low shrubby and herbaceous species are the same,
with a number of new ones, most of which are characteristic of the
nearby forest. The list follows:

Micromeria chamissonis. Rosa californica. Chlorogalum pomeridianum.

Dominant. Trientalis europsea latifolia. Symphoricarpos racemosus.

Aster radulinus. Domi- Gymnogramme triangularis. Galium californicum.

nant. Pedicularis densifiora.

There is considerable litter and a moderate amount of humus.
Bed rock is exposed in a few places.

38 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

Station 7.

Descending still farther, the slope becomes steeper and we pass
through a transition zone into the forest which clothes the lower
part of this north-facing hillside. It is not an extensive area, the
altitudinal distance covered by it being only 40 meters, but it is
nevertheless well developed and typical in every way. Two stations
were located in it. Station 7 is in a representative spot 12 meters

Fig. 5. — Quadrat at station 6, 5 meters square. Note likeness
to station 3 (fig. 3).

above the bottom of the ravine. Since trees are the dominants
here, it was necessary to increase the size of the quadrat to 10 meters
on a side. This quadrat, therefore, equals in area 4 of those in other
localities. A summary, as far as the dominants are concerned, is
given (fig. 6):

Clumps. Stems.

Quercus agrifolia 6 6

Arbutus menziesii 2 8

^Esculus californica 2 2

The specimens of Quercus and JEsculus are all single-stemmed
individuals. Those of the latter species are of little importance,
being but 3.8 and 6.3 cm. in diameter. The oaks range in diameter
from 3.8 to 36 cm., four being 10 cm. or larger. The eight Arbutus
stems are in two clumps of large stump sprouts, a very common
habit of the species. The diameters of the individual stems range
from 10 cm. to 18 cm. The area is shaded and controlled by the

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 39

three or four largest oaks and the two Arbutus clumps, together
with other trees outside the limits of the quadrat.

If we should make a more general survey of the locality, we
would find that the quadrat represents average conditions faithfully,
except that another tree species, Umbellularia calif ornica, occurs
along the ravine-bottom and in side gulches. None of the trees
are tall, the mature of all species ranging from 7 to 12 meters in

Fig. 6. — Quadrat at station 7, 10 meters square. The symbols
surrounded by double lines indicate trees of the domi-
nating stratum; the others represent the shrub stra-
tum; the herbaceous stratum was not charted.

height. The oaks have mostly single trunks, those toward the
bottom of the slope being much larger than those above. One
specimen, 10 meters above the ravine-bottom, had a diameter of
8 m. The madronos mainly show the clump habit. Seedlings of
oak occur, but are very rare. No madrono seedlings were found.
The shade is nearly everywhere dense.

Unlike the chaparral, the oak-madrono forest shows distinct
vertical zonation or layering. The dominants are the trees, and
beneath them there are two subordinate strata, one composed of
tall shrubs, the other of low shrubs and herbs. The first is few in
species, but one of these, Rhus diversiloba, is abundant. Others, all
relatively unimportant, are Berberis pinnata, Holodiscus discolor
aricefolius, Rosa calif ornica, and Dirca occidentalis. With these
should be included a few frequenters of disturbed areas, such as

40 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

Ceanothus sorediatus, Eriodictyon californicum, and Diplacus glu-
tinosus, which occur in small numbers in the few openings; also,
occasional wanderers from the chaparral; and finally, the young
individuals of the dominant trees. The same quadrat includes the
following shrubs:

Rhus diversiloba 59

Holodiscus discolor 2

Quercus durata 1

The herbaceous and ground-shrub vegetation is abundant, both
in species and in individuals. The following list is nearly complete,
the three most important being numbered in order of abundance:

Gymnogramme triangularis. Smilacina amplexicaulis. Cynoglossum grande.

Adiantum jordani. Ranunculus hebecarpus. Micromeria chamissonis (2).

Aspidium rigidum ar- Dentaria integrifolia. Pedicularis densiflora.

gutum (1). Saxifraga californica. Galium aparine.

Melica imperfecta. Tellima heterophylla. Galium californicum.

Luzula comosa. Potentilla glandulosa. Symphoricarpos racemosus (3).

Fritillaria lanceolata. Psoralea physodes. Hieracium albiflorum.

Trillium sessile gigan- Trientalis europsea lati- Achillea millefolium.

teum. folia

As

c

5

1

a' 5

A<
A 4

Rock

D

A 3

DAs

Fio. 7. — Quadrat at station 9, 5 meters square. Dominance of
comparatively few individuals of Adenostoma.

At all times of the year this vegetation is in marked contrast to
the bareness of the ground beneath the chaparral. In the spring
the greenness and luxuriance are especially striking, when such
plants as Fritillaria, Trillium, Smilacina, Tellima, and Cynoglossum
are conspicuous. Litter is naturally abundant, and the soil is well

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 41

supplied with humus therefrom. The occasional rock outcrops
bear a fair amount of moss vegetation.

Station 8.

This is similar to the last, but is close to the ravine bottom.
Umbellularia californica is very important, growing in dense groups,
which cast almost complete shade. Lower vegetation is almost
absent in such places (although a healthy seedling of Umbellularia
was found), but elsewhere the herbaceous element is like that
described above.

Station 9.

A sudden change of habitat, due to the sharp V-shaped bottom
of the ravine, brings about a transition in vegetation that is even
more abrupt than that between stations 3 and 4. Facing the
oak-madrofio slope and beginning in full development at the very
bottom of the ravine, we find a nearly pure growth of Adenostoma.
The shrubs are larger here than in any previous station dominated
by that species, averaging 1.5 meters in height. The living and
dead branches are so densely interlaced that progress even on hands
and knees is almost impossible. The following summary of the
quadrat (fig. 7) gives a satisfactory picture of the composition:

Clumps. Stems.

Adenostoma 29 175

Ceanothus sorediatus 1 1

Total 30 176

The one specimen of Ceanothus sorediatus overtops everything
else. The shade is dense, considering the fact that Adenostoma is
responsible for most of it. Ground-cover is very sparse and consists
of Aster radulinus, Chlorogalum pomeridianum, and occasional small
plants of Diplacus glutinosus. Litter and humus are scanty and
the steep slope shows frequent rock outcrops.

Station 10.

This is situated on the gently south-facing slope, 130 meters south
of the last. It was chosen for special study because it was in the
only chaparral area where the soil was such that excavation for
water-content determination could be made in it to the desired
depth. The fact that the vegetation has been slightly disturbed in
consequence of its proximity to an old road does not vitiate its
usefulness for this purpose. Adenosto??ia is dominant and large.
Arctostaphylos and Quercus durata are both frequent, and Q. wislizeni,
Heteromeles, and Prunus ilicifolia also occur. Because of disturbance,
Eriodictyon calij 'ornicum, Helianthemum scoparium, and Diplacus
glutinosus are frequent, with quite an assemblage of small wet-
season annuals. Litter and humus are scarce. The uniform sandy
soil grades evenly into the bed rock, becoming too hard for excavation
at a depth of a little more than a meter.

42 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

THE HABITAT.

The foregoing description, assisted by the profile (fig. 1), makes
it plain that superficially the distribution of the various communities
depends altogether upon the topographic factor slope. This of
course is a complex of simpler factors and the present study therefore
resolves itself mainly into an analysis of those factors whose variation
depends upon differences in direction and angle of surface inclination.

At this point a summary of the observed angles may well be
given.

Table 2. — Direction and angle of slope.

Station.

Vegetation.

Direction.

Angle.

Station.

Vegetation.

Direction.

Angle.

No. 1
2
3
4
5

Adenostoma. . .
Adenostoma. . .
Arctostaphylos
Adenostoma. . .
Adenostoma. . .

S

N
S

s

8

32

6-14

5

No. 6

7

8

9

10

Arctostaphylos

Forest

Forest

Adenostoma. . .
Adenostoma. . .

N
N
N
S
S

30-34
37
42

23-38

7

THE SOIL FACTORS.
Physical Character.

Samples were taken from a selected series of stations — Nos. 2, 3>
4, 6, 7, 10 — these being the ones used in the study of water-content-
From the first four, two each were taken, from 10 cm. and 30 cm.
depth respectively. In stations 7 and 10 three depths were sampled:
10 cm., 50 cm., and 100 cm. All of these correspond with the
depths used in moisture determination in each habitat. Since no
differences of moment were discovered between samples from a
single station, these have been combined and averaged in each case.
The analyses were made by Mr. Alfred Smith, of the University of
California, through the kindness of Dr. Charles F. Shaw. In the
first part of table 3 the five grades of sand and gravel have been
condensed into two.

Table

3. — Mechanical analysis.

Station.

Gravel.

Sand.

Silt.

Clay.

Gravel +Sand.

Silt+Clay.

2

2.30

55.71

12.66

28.96

58.01

41.62

3

2.01

78.82

2.42

16.31

80.83

18.73

4

5.58

72.08

2.01

19.92

77.66

21.93

6

1.52

76.05

2.04

20.09

77.57

22 . 13

7

4.61

73.21

12.54

9.34

77.82

21.88

10

1.93

81.73

7.88

8.02

83.66

15.90

It is seen that the soils in all but station 2 are closely similar,
with the proportion of gravel and sand running from 77.57 to
83.66 per cent. Some, according to Dr. Shaw, would be classed as
light sands and others as sandy loams. Their similarity would be

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 43

assumed from field observation, the only noticeable difference being
that some are a little more coherent than others. The soil of station
2 is strikingly different, having much less of the coarser grades and
a correspondingly larger proportion of silt and clay, especially the
latter; consequently it is sticky and heavy when wet, is unfavorable
to water movement (shown by the fact that water stands in holes
for many hours after a rain), and is very hard and almost rock-
like when dry. Though still belonging to the sandy-loam class, it
tends strongly toward clay. This physical character results in
quite a different set of soil-moisture values, as will appear later,
and it is therefore not surprising to find the difference in the vege-
tation which has been noted— a difference, however, entirely in
relative luxuriance. The importance of the mechanical analysis to
the present study is that it demonstrates that the soils of most of
the localities studied do not differ in physical nature (excluding the
humus portion), and that therefore this factor as a determinant of
water-content and of vegetation has been excluded, since it is not
a variable. The stations in which no analyses were made, from
field observation should be grouped as follows: station 1 like sta-
tion 2, 5 like 4, 8 like 7, and 9 like 10.

The same samples were submitted to Dr. Charles B. Lipman, of
the University of California, for determination of humus. The
depths are those noted in the last section. T indicates less than
0.1 per cent.

Table 4. — Average humus content.

Station.

Vegetation.

At 10 cm.

At 30 cm.

At 50 cm.

At 100 cm.

Nos. 2,4, 10
3, 6

7

0.11
0.34
1.97

T
T

T
0.85

0.66

The importance of this reaction-factor in the surface layer is
seen, and its penetration to a considerable depth in the forest.

The Soil Moistuke.

The determinations of water-content were made in two series.
In the 'first and more important, two stations were utilized, Nos.
7 and 10, the object being to contrast the conditions in chaparral
and forest. Determinations were made weekly for a period of one
year, and the series was later resumed for a few weeks for a special
purpose. In the second the stations selected were the six used in
the evaporation experiments and the same from which the samples
for mechanical analysis were taken. This series ran parallel to the
evaporation work, determinations being made at intervals of four
weeks from June to April.

44 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

Series I.

Station 7 is an altogether favorable spot for the study of the forest
soil. Station 10 was selected to represent the chaparral because it
was the only one having a soil in which it would be possible to
excavate to the desired depth. For the study of conditions surround-
ing the deep-rooted chaparral species, a knowledge of the water
conditions in the deeper soil-layers seemed essential. Even here the
deepest excavation possible did not approach the region reached by
the longest roots, which, as will be shown (p. 89), penetrate the
undecomposed sandstone to a considerably greater depth. Extreme
shallowness of soil is general over the chaparral area in which the
studies were made. It was felt that this advantage in depth of soil,
which station 10 possessed, compensated for certain disadvantages —
lack of slope contrast to station 7, the angle being only 7° southward,
and slightly disturbed condition, due to proximity to an old road
and the edge of the chaparral area.

The series was begun January 21, 1913, and continued for a
total of 54 weekly observations, ending January 26, 1914. Occa-
sionally a visit was delayed for a day or two by storms or other
causes, especially in the winter of 1913-14; otherwise the program
was carried through without a single break or mishap of any kind.
Samples were taken in each station at three depths: 10, 50, and 100
cm. Excavations were made with a spade and samples were col-
lected in numbered and weighed glass bottles. The sample was
obtained, whenever conditions permitted, by pushing the mouth of
the bottle into the freshly exposed soil-surface. When this became
impossible toward the end of the dry season, especially at the lowest
level, lumps of soil were cut out and pressed into the bottle, which
was immediately tightly corked. The first weighings were always
made within a couple of hours after digging the samples and were
carried out to 0.01 gram, which is probably greater accuracy than is
necessary in such determinations. The soils were dried at a tem-
perature of 105° C, and the water-loss was computed upon the
basis of dry weight.

The results have been plotted together in figure 8. The first
winter was a very dry one, the total precipitation at Palo Alto
being only 19.49 cm. Assuming a proportional relation at Jasper
Ridge, the rainfall at the latter locality should have been 32.5 cm.
The winter of 1913-14 was in great contrast to the preceding one,
the precipitation at Palo Alto being 62.15 cm. and at Jasper Ridge,
by actual measurement, 103.71 cm.

Taking first the three curves representing the soils of station 10,
in the chaparral, we observe the following facts: The effect of
rainy and dry seasons is the most conspicuous feature. The highest
point reached in the first rainy season was 13 per cent, in the second

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 45

17.2 per cent. The absolute minimum was 0.5 per cent on October
31. We next observe the gradual dropping of the three lines,
beginning with the end of the spring rains, continuing at a constantly
decreasing rate through the dry season of summer and fall, and

^2*

* ■

s,

i^

£

-

,

.- 1

jj

■*•

■:i

r

■

. fVJ

— 1

-.

^

ya

r

'■I

-.

•4

--~.

w

•

?

-^

i-

s

(U

__

•-.

-

~J

~ h

"-

—

--

--,

_

3*M

° *

Z -

=3i

^ B

' •■

& if±

•t

1

w

|_ M

■'

o -

'. i \

-

x

**. a. - ■

r- <u

3 E : --
U1 U '

\i

] ,

!/

!r £

3

{

il?

7

10 -

I

/>'•

} rf '

1

1

i'rh

*

° r-

in

1 "

1 |

1 .
1 .

1 fl : ;

O -

' X:

^ R]

(

1

/ 1*3

3 *

X

^ -p

o S" : :
- . -u

±J t v :

in u " i

r 9S.8

,f:j

■^

y;/>: !

-> -

TM ■

i

tn

1

,1 i : i

U RJ

Zw

\ ■• • i

v :

3 -

< 4\

1

'■ ) i:

SI

< w

55

(_

ti-
ro

X

L-

.•'

1

,

1 ;

4

,

-y

'••

'•'

»/

f

;

y,

0. N
< in

-

^

,'.

/

! :

,. •

■ :

'

a. !0

«-

-'

'

,-

-.

; j

< a

2 -

a

»

^Z

~ j ?~-

-T:

-

/

r'

•

'

in

f5

^

s

'■

CD =

--

-

-~"

■

\

lL "

"■

~~

rtT

3

">>

/

/

»<:

o> n -

-

i

1 c

3 I
f c

s
MC

n c

1 r

a:
)i j

3 L

i c

1

V

| i

j r

31"

1 1

3 II

11
=11

3 C

N"
D3

3 1

H,

*>
j

r
f

- ■

J c

j b

fl

u c

1)

1

10

M

3i

A

j

n

o r
4

M3

II ■

t

1

d

j
' 3

- <

3
1J.

91

D

- t

o in
IO!

-i

-i rj

Fig. 8. — Soil moisture at weekly intervals for one year, in chaparral (station 10) and broad-
sclerophyll forest (station 7), at depths of 10, 50, and 100 cm. Jasper Ridge. The
wilting coefficients for the 10 cm. and 100 cm. depths have been added for each
station.

46

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

thus approaching a condition of constancy. The sudden rise with
the first autumn rain is the most striking single feature. Comparing
the three lines, we see that the 100 cm. level had the highest water-
content through most of the year, and the 10 cm. level the least.
The order was suddenly reversed immediately after rains, the
most striking case being at the beginning of the rainy season of
1913-14. The first rain occurred on October 31, and was a sudden,
heavy shower, followed by others at frequent intervals during the
month of November. From October 31 to November 6, inclusive,
the precipitation at Palo Alto was 3.5 cm. Figure 8 shows that on
November 7 the water-content at 10 cm. depth had risen from 0.5
per cent to 5.2 per cent and on November 14 to 10.2 per cent.

The lag in the response of the lower levels is interesting. At
50 cm. depth no effect was observable until November 22, and the
increase when it came was less sudden. The water had not reached
100 cm. depth on December 5 and the rise thereafter was still more
gradual. The slow penetration of the rain-water is not due to
run-off, since the slope is only 7 per cent, but doubtless, in the main,
to the air-filled condition of the soil. Table 5 shows roughly the rate
of penetration in the two stations during this period. No important
difference between the two stations is seen in this respect, and the rate
of penetration was approximately uniform as far as it was followed

Table 5. — Rate of penetration of first rains.

Depth of penetration in cm.,
first rain October 31.

Station 7. Station 10.

Nov. 14
Nov. 22
Dec. 13
Dec. 22

31 35
50 to 60 58
70 to 100+ 90
75 100 +

Returning to station 10, we note further the tendency toward
convergence of the three lines as the dry season progresses. The
sudden jumps of July 13 and August 31 in the 100-cm. line are
difficult to explain. They may possibly be due to encountering
small masses of soil where the clay-content was above the average.

We may draw the following conclusions concerning the soil-
moisture conditions in this particular chaparral station during the
period studied. During the actual continuance of the rainy season
there is an abundance of water at all depths. This was true even
in the unusually dry winter of 1912-13. As to the following very
wet season, the soil at station 10 was frequently so thoroughly
saturated that the walls of the excavations continually collapsed
into an almost soupy mass, making it difficult to obtain fair samples

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 47

at the lowest depth. During the spring and early summer of 1913
the water-content steadily decreased, and by mid-July it had reached
a point beyond which it could not drop much farther, the samples
when gathered appearing as incoherent air-dry sand. From then
till the first rains, three and a half months, there was a period of
extreme deficiency, which is not exceeded in severity by conditions
as recorded in the desert of Arizona (57). The lag of several weeks
in the response of the lower strata to the first rains shows that their
advent does not close the dry season, so far as the soil is concerned.
In the forest station the general form of the curve is the same,
but there are nevertheless important differences. The maxima are
uniformly greater than in the chaparral. Comparing the three
depths, we find the relative positions of the 10-cm. and 100-cm.
lines the reverse of those in the chaparral, the 10-cm. depth showing
nearly always the highest water-content and the 100-cm. depth the
lowest. Obviously this is to be traced to the effect of the humus on
water-retaining capacity. The lines, too, during the rainy seasons,
are much more widely separated than in station 10. During the
dry season, however, they converge until they are as close together
as those of the chaparral. The response to the first rains in the
deeper levels shows a more noticeable lag than appears in the chapar-
ral, the first certain increase at 100 cm. depth appearing on December
26, eight weeks after the earliest shower. Comparing equal depths
in the two stations, we find that at 10 cm. the forest line is uniformly
higher than that of the chaparral, the reasons being shade and
humus. The maximum difference between them during the rainy
season is 14 per cent, and the average for the rainy seasons ap-
proximately 8 per cent. During the progress of the dry season
the lines converge, the average difference for the last three and
a half months being 3 per cent. At 30 cm. depth the forest line is
still well above that of the chaparral, but the differences are uni-
formly less in both wet and dry seasons. At 100 cm. no such
constant differences are seen. During the first winter the chaparral
soil at this depth was wetter than the forest; during the dry season
and the second winter the lines cross and recross without revealing
any differences of importance.

The points of greatest general import are: (1) the rather striking
difference in water-content between chaparral and forest during the
rainy season; (2) the convergence of the six water-content lines to
minima which at the end of the dry season are not far apart.

The depletion of the water-content in every case is due to three
causes. Drainage of the gravity water through the soil takes place
with great rapidity during the wet season, which is proved by the
fact that the ravines of Jasper Ridge contain running streams for
considerable periods. After the rains have ceased, however, it can

48 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

not be long before all the gravity water will have been removed from
the upper meter of depth, so that this cause will no longer be oper-
ative. The second cause, evaporation directly from the soil, and
the third, removal by plants, will continue throughout the dry
season. Evaporation from the soil is undoubtedly higher in the
chaparral habitat than in the forest, and yet the gradient of depletion,
as shown by the graph, is decidedly steeper in the latter. This
seems to show that the most potent cause of water-content depletion

ZW2

Oft:
< *-

K hi I

1913

SEPT. OCT. NOV.

12 19 26 3 '0 17 24 31 7 14. 22

1914
SEPT.
17 24- I

OCT. NOV.

8 15 22 30 5 12 20

! .

i., i

i i

T — 1

/ {

i

i
t

i-i

7

/•

j

/ 1

/ i

' i

-

j i

V

f__~7

S

:•/.<

•*s

...-•

-iiv -

:.*:..

"-* ']•

=^^=

^_ -

1

—

—

—

~" r

■

oq:

rid

20
I 5
I
.5

U .0
19
18
17

SE'6

2,5
UJ

S l4

>- 13
DC
O 12

o"

I- 10

z

7

— .r

■ '"' x --'' \ -'''7

^J\t_^^_

I '•/ \

j/*- — ^jr 7^\

" — '•••:£= = — S^E"^

■"** / \^ " "*"

/ ' i

Fig. 9. — Soil moisture at end of dry and
beginning of wet season, fall of 1913.
Jasper Ridge. Wilting coefficients as
in fig. 8.

Fig. 10. — Soil moisture at end of dry
and beginning of wet season, fall
of 1914. Jasper Ridge. Wilting
coefficients as in fig. 8.

is the vegetation itself, since where the plants are largest and most
abundant, there will be the greatest drain upon the water-supply,
which statement exactly fits the forest in this case. Measurements
given in another connection (p. 112) show that the rate of transpira-
tion in the same species growing in both habitats is much higher in
the chaparral habitat (in the case of Adenostoma, forest : chap-
arral : : 1 : 1.92; in the case of Arctostaphylos, 1 : 1.69). However,
it is reasonable to assume that the far greater bulk of transpiring
vegetation in the forest, comprising trees of large size and luxuriant

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

49

herbaceous vegetation, will much more than counterbalance the
greater rapidity of the process in the chaparral.

The graph (fig. 9) shows dry season conditions after a winter
of deficient rainfall (1912-13). The following winter was a very wet
one, and it seemed worth while to resume the soil-moisture deter-
minations for a few weeks during the critical period of the dry

sis

££;?, £]^ri8 <n<Dr * , *>« n *<'><v»-offj«>r»<oift*WN-o

(M M OJ N N N N

J.H9I3AA A^a AO J.N3D «3d '3tin.LSIOIM "1IOS

50 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

season following. The two graphs side by side (figs. 9 and 10) give
the minima shown in table 6. This shows that there are very
perceptible differences in water-content at the critical periods

Table 6.

Years.

Chaparral.

Forest.

10 cm.

50 cm.

100 cm.

10 cm.

50 cm.

100 cm.

1913
1914

0.5

1.1

1.1
1.1

17
5.9

3.7
5

1.9
3.7

1.8
3.9

following winters of deficient and abundant rainfall, and that these
differences appear especially in the deeper soil-layers. In the
chaparral, at 100 cm. depth, the minimum percentage after a wet
winter was more than three times that of the preceding season;
at the corresponding forest depth the percentage was more than
twice that of 1913.

Series II.

This series was carried out at the same six stations that were
used for the evaporation studies, and during the same period (June
15, 1913, to April 17, 1914). There were thus provided three
examples of Adenostoma chaparral (Nos. 2, 4 and 10), two of
Arctostaphylos chaparral (Nos. 3 and 6), and one of forest (No. 7).
The samples were taken at intervals of 4 weeks, and in stations 2,
3, 4, and 6 at two depths — 10 cm. and 30 cm. In stations 7 and
10 the figures from Series I were used, the 50-cm. depth being con-
sidered as parallel to the 30-cm. depth in the other stations. In
figure 11 the results for the 10-cm. depth are compared.

The most striking feature is the line representing station 2, the
highest in water-content and yet the poorest in vegetation. The
explanation of this apparent paradox is found in the mechanical
analysis and wilting coefficient. The sudden changes in water-
content without apparent reason — e. g., the sharp crest on July 13
and the sudden drop after March 4 — are perhaps due to local soil
differences, small masses of unusually high or unusually low clay-
content. The other two Adenostoma stations are consistently the
lowest and the forest is the highest. Between forest and Adenos-
toma chaparral are the two Arctostaphylos stations. In figure 12 all
stations of each type, both depths, have been averaged, station 2
being omitted. The daily precipitation for the period is also added.
The Adenostoma community is decidedly the lowest, and the forest
is highest most of the time, though the Arctostaphylos community is
not far below. The close approach of all the lines during the dry
season and their separation during the wet season (seen in Series I) are

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 51

evident here also. In general, Series II confirms the results obtained
in Series I and adds data for the Arctostaphylos community, which is
seen to be intermediate in water-content between the other two.

N

t
i

v

\
>

O

(VJ

f

1

< -i

1

=5

^

•

/

«Z,n

H

/

/

/
/

^<M

\

\

V

" 1 '

v

! 1

U <^
D

a

1
1

\

1

i

■^^

**

^

^

%s

^

.^

i

o

i
i

Ndenostoma
:tostaphylos

Ifl

u

L

U.

/

/

1

uj *

O

DO

<"

i

<

5„

— * ifl

i

1;

/

/

omonoiAOinoino

SH3X3VNIXN3D
'N0iXVXldO3Hd

XH9I3M AMQ jOXN3DH3d'3dnxSlOKN "1109

JB

-

«~ .2

C3 o,

M-O

O .C
X (M

C -

£ 2

3 S

Water-retaining capacity. — The reasonable way to compute water-
retaining capacity is upon the basis of dry volume of soil. The
same is true of water-content, but as yet it is not feasible to make
moisture determinations in extensive series on the basis of volume.
For the sake of uniformity, therefore, it is convenient to express the
water-retaining capacity in terms of dry weight as well as of dry

52 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

volume. The method employed was the one which has been in
general use of late. Circular pans were made having bottoms of
perforated metal, with sides 1 cm. high and of such an area that the
capacity, when the soil was smoothed off level with the top, would
be 100 c. c. The bottoms were covered with cheese-cloth. The soils
were all prepared — pulverized and compacted — as nearly as possible
in the same way, and the pans, filled level full, were stood in shallow
water to permit absorption. After saturation was complete they
were allowed to drain until all the gravity water had passed off.
The percentage of water retained was then calculated upon the
basis of dry weight and also of dry volume. In each case the average
of five tests was used. The shortcomings of the method are fully
realized, prominent among them being the unnatural condition of
the soils when tested. The results, however, although they do not
give a true picture of the behavior of a given soil under natural
conditions, do give data whereby fair comparison may be made
between different ones. Samples were taken from the three depths
in stations 7 and 10 for comparison of conditions in forest and
chaparral, and from station 2 because of its peculiar qualities.

Table 7. — Water-retaining capacity, percentages.

Station.

At 10 cm.

At 10 to
30 cm.

At 50 cm.

At 100 cm.

Average
all depths.

Ratio.

By dry weight:

No. 10, chaparral. .

No. 7, forest

No. 2

20.6
30.5

38.5
49.3

46.4
71.4

17.9
24.9

36.7
47.3

21.4
16.3

39.4
32.5

20

23.9

46.4

38.2
43.0
71.4

1

1.20

2.32

1
1.13

1.87

By dry volume:
No. 10, chaparral. .

No. 7, forest

No. 2

Considering the dry-volume figures, which give the truest picture
of the soil qualities, we note that there are no differences of im-
portance among the various depths in station 10. The high water-
retaining capacity in station 7 at 10 cm. depth is plainly due to the
large humus content. The high figure for station 7, 50 cm. depth,
as compared with the low one for 100 cm., is difficult to explain.
Mechanical analysis offers no clue; nor does humus content, since
50 cm. and 100 cm. are more like to each other than to 10 cm. depth.
Station 2 shows striking originality in this feature as in others, and
the reason is plain — the high percentage of clay and silt. Sum-
marizing, we find that the comparative water-retaining capacities
of stations 10, 7, and 2, representing chaparral, forest, and a par-
ticularly unfavorable chaparral station, may be expressed by the
ratio 1 : 1.13 : 1.87.

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 53

A second series of determinations was made at a later time. The
percentages are somewhat higher, but the relative values are very
similar in most cases, which is the important thing. The series
given here is the more valuable, since the soils used were composites
of many individual samples.

Wilting coefficient. — The method developed by Briggs and Shantz
(15) was followed, and the soils to be tested were brought from Jasper
Ridge to the University of Minnesota. The experiment was carried
on in the university greenhouse, in a room where the relative
humidity averages 30 to 40 per cent during the middle of the day.
An atmometer was kept in operation close to the jars during most
of the period of growth and wilting. The average daily evaporation
recorded, reduced to the usual standard, was 10.2 c. c. The extremes
were 14.7 c. c. and 3.1 c. c. Glass tumblers were used for containers
and a constant soil temperature was maintained by immersing them
almost to the top in slowly running water. Kubanka wheat,
obtained through the kindness of Dr. Briggs and Dr. Shantz, was
used as the indicator plant.

The constant termed the "wilting coefficient" has been variously
interpreted. The latest and most reasonable conception, due to
Shull (82) and Moore (65), is that it is purely a function of the
soil, "a point in the water-content of the soil at which the water
practically ceases to move along the film (on the soil particle), no
matter how sharp the gradient at the edge of the film," and not a
matter of "balance between the 'back-pull' of the soil and the pull
of the plant." If this be true, one plant is as good as another for
indicator use, and Kubanka wheat will furnish just as serviceable
results as would Adenostoma. This is fortunate, since difficulty has
been encountered in bringing about germination of the chaparral
species under artificial conditions. The one chief essential to
success seems to be the maintenance of uniform environmental
conditions during the period of the experiment, especially those
affecting the evaporation-rate. It must never be forgotten that
the wilting coefficient is the ultimate point at which plants must
wilt, but that under certain conditions wilting may take place
before that point is reached. Its usefulness in the investigation of
different soils with regard to their effect upon plant life is obvious.

Comparison was made primarily between the soils of two stations,
Nos. 7 and 10, representative of forest and chaparral respectively.
In each case two depths were considered, 10 and 100 cm. Station 2
was added because of the interesting singularity apparent in its
other characteristics. The figures in table 8 are each an average
of seven unimpeachable determinations.

The figures are about what one would expect after study of the
physical character of the soils and their water-retaining capacities.

54 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

Station 10, almost a pure sand, has a very low wilting coefficient,
with no difference of moment between the two depths, since the
surface layer has but a very slight admixture of humus. In station 7
the effect of surface humus in raising the wilting coefficient is very
evident. In station 2 the large percentage of silt and clay is respon-
sible for a strikingly high figure.

Table 8. — Wilting coefficient in stations 10, 7, and 2.

Depth.

Station 10.

Station 7.

Station 2.

10 cm. . .

100 cm. . .

p. ct.
1.5
1.7

p. ct.
4.5
2.3

p. ct.
16.1

It has been customary to draw a line to indicate the wilting
coefficient upon the graph representing the seasonal march of soil-
moisture. Accepting the interpretation of the wilting coefficient
presented by Shantz and Moore, we are justified in drawing this
line, and in assuming that, when the water-content line falls below
that of the wilting coefficient, the vegetation can extract no moisture
from the soil. Referring to figure 8, then, we find that in station
10 (chaparral) there was at 10 cm. depth no available water from
July 1 to November 1. At 100 cm. depth the water-content line,
though it is not far above the wilting coefficient, actually touches
it but once. In station 7 (forest) conditions at 10 cm. depth were
only slightly better than in station 10, water-content and wilting
coefficient lines both being relatively and equally high. The period
of unavailable water was shorter by the first two weeks of July.
At 100 cm. depth conditions were actually less favorable than in
station 10, since three times the water-content line descends below
that of the wilting coefficient. With regard to the relation between
water-content and wilting coefficient, it would seem that there is
no striking difference between stations 10 and 7; that soil-moisture
conditions, as indicated by these two factors, are about equally
severe during the critical period.

The dry season of 1913 was of unusual severity. That the
conditions recorded are extreme is indicated by figure 10, presenting
the soil-moisture in relation to the wilting coefficient at the critical
period of 1914, which followed a very wet winter. Except for
station 10, at 10 cm. depth, the water-content was everywhere
well above the wilting coefficient, and the abundance of available
water at 100 cm. depth in station 10 is noteworthy. Very likely
the average condition is somewhere between the extremes of 1913
and 1914, but the special importance of seasons of unusual severity

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 55

(like that of 1913), which recur every few years with tolerable
regularity, is fundamental.

In station 2, the constantly high water-content is rendered in-
effective by the high wilting coefficient. At 10 cm. depth, during
the period from June 15, 1913, to April 17, 1914, the soil furnished
available water only from November 1 to April 1. It seems un-
necessary to seek further for the explanation of the poverty of the
vegetation in that station and its vicinity.

The Soil Temperature.

In making determinations of soil temperature, the simplest
method seemed the best — the plunging of a chemical thermometer
into the freshly cut surface at the time of taking water-content
samples. During the dry season it was found necessary at the
lowest level to bore a hole with a small auger. The observations
were made in the two stations of water-content Series I, at all
depths used in that series. They were begun March 25, 1913, and
with a few minor interruptions continued to January 26, 1914, the
closing date for the soil-moisture work. During much of the period
the temperature of the surface was also observed, in shade in
station 7 and in sun and shade in station 10. In these cases the
bulb of the thermometer was covered with a thin layer of soil or
litter. The results are plotted in figure 13.

The first point to note is that during the dry season the soil
temperatures in the two stations are widely different, the three
lines representing station 10 being all of them regularly below those
of station 7. With the arrival of the wet season they all descend
and at the same time converge, the differences between them almost
disappearing. This is exactly the reverse of the soil-moisture's
behavior, where convergence takes place in the dry season. It is
due of course to the greatly increased water-content of both soils.
The surface temperatures, naturally the highest in each station
through most of the period, show the same tendency, and in the
case of station 7 the line actually drops below those representing
the subterranean temperatures.

Comparing the lines of a single habitat, we find that in station 7,
during the dry season, the temperatures vary inversely with the
depth, the order being, from highest to lowest, surface, 10 cm.,
50 cm., 100 cm. During the wet season the order is roughly the
reverse, the order most of the time being 100 cm., 50 cm., 10 cm.,
surface. In station 10 the case is the same, except that the shaded
surface and 10 cm. lines do not greatly differ, because of the lack of
cover, and in the wet-season portion the lines do not hold constant
relative positions. In both stations the greatest fluctuations occur
at the surface and at 10 cm. depth, and the greatest uniformity at
100 cm. depth.

56 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

THE ATMOSPHERIC FACTORS.
Light.

On August 17, 1917, a series of observations was made to determine
the relative light values at the different stations. The apparatus
used was the Clements photometer. The observations were made
as rapidly as possible in the course of a trip over the trail from
station 1 to station 10. The first was made at 2 h 30 ra p. m. and the

r 3 ^

*<*

%

s

ss

,^

*

g

|TS

.

—

If

^ i V

_^ ^ o 9 o

^.-

.-

A

S-*

"--

--

vj C-J
Q rO

in

i

4

7

r

A

f*

/

M

in 1

1

1 •
1 -

i ;

I :

Ss

o 1

2 ^

/,

'J

£

--■

•^

c^v

yv

■r

/

/

■' N,

T7

*

.,-

"""

i.

H

u
o °

/

/

^'

Vt " f

,. —

--

"•—

/

J 1

—

—

-..

-..

_ .

•v..

f i Lr

10

.--

—

—

■—

_^Z

xif 4 *

J)

^ 01

-.

//

"^M .>

x5

Q. ~

\

v^

s&

<vl

*/> co

/'

N

b-

■-

-—

—

r ■

/ :

10

.

—

■•"

[ —■«

/

•vi

—

—

-__

X

'4

i^

i '

SJ/.

3

<2

--

—

— '

-<f

/ft

--

...

$"

~-'<

...

j\

\

\\

> O

1 3

r-
/

-> -
en

i — ■ i

L

\' v *

OJ

to ' ■ » " "

-P 4) <U U

co u o •
w , cu.ro F . :

lj_ <+- h - ■

-= t_ t_ <-* . 1

ro -, t t ' o

['•

UJ m

2 in

/

^

\

/

^2

■-

-^

\

~~L

1

k

/

/!

H

i

/

CD

*

"j^

k

> O

< t\J

2«

<

pi"

>

co : |

°- 1

TO

-c ! !
u 1 1

"^

-— .

•O

"--

"V

k \

"t"

.7

S

s

V.

S

"h

_i n

*

^ti

rfS

\"

•V

?■"

"**"

-.

\

V

*

X.-

.•1

<■

/

r

V

k

!XI

"5 2

^

'/

1/

) u

) a

) ^

) u

<

r
■ «

c

) a
r P

) u
1 p

3 <j

- (\

)

J c

1 *

1 !i

! S

J r C

! 2

-> a

) <

r>

J c

«

n

^

■ t»

J o

^ 01

s ^
8 a

So

O 73

- o>
3^

"O —

> 9-

- o

CD (i
-»^ OJ

01

^1 03
CD O

OJ t-

w c

a. 2

3aVdOLLN33 S33y93Q l 3ynxva3dlAJ3X TI09

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 57

last at 3 h 15 m . Those toward the end are therefore comparatively
a trifle too low. In the computation the highest value obtained
(at station 5) is taken as unity and the others are reduced to decimals
in proportion. In the observations of sun intensity the instrument
was pointed directly toward the sun. In the spotty shade of the
chaparral it was moved about continually during the exposure, so
that an average figure was obtained.

Table 9. — Light intensity at stations 1 to 10.

Stations.

Sun.

Shade.

Stations.

Sun.

Shade.

Summary.

Sun.

Shade.

Proportion.

No. 1

0.94

0.20

No. 6

0.69

0.10

Adenostoma. . .

0.84

0.17

100 to 20

2

.75

.17

7

.56

.07

Arctostaphylos

.69

.07

100 to 10

3

.69

.04

8

.05

Forest

.56

.06

100 to 11

4

.75

.21

9

.11

5

1.00

.23

10

.75

.07

The great differences in the direct sunlight intensities are appar-
ently due to differences in amount of reflected and diffused light,
the chaparral stations, with much bare light-colored soil naturally
being the highest, and the forest, with dark soil and much vegetation,
the lowest. The shade differences are less than might be expected.

Temper attjee.

In the matter of atmospheric temperature there is an unfortunate
shortage of data. Isolated observations of so variable a factor,
even though numerous, are of little value. The thermograph is the
only satisfactory apparatus for habitat study, and no thermographs
were available for the work. Some knowledge of comparative
temperature values during the dry season may be gained from two
series of observations (table 10), obtained during trips over the
trail connecting the stations. The first was made during the middle
of the day, the second between 3 and 4 p. m. The readings were
made from the dry bulb of the cog psychrometer in the process of
measuring relative humidity.

The outstanding feature is the uniformity of temperature. No
differences of moment occur between any of the stations, nor be-
tween sun and shade. The latter would, of course, not be true if
the sun had been allowed to shine on a stationary bulb. The
probable effect in such a case is roughly indicated by the high sun
temperatures recorded in the surface lines of the soil temperature
graph (fig. 13). As the air readings stand, the shade temperature
in the Adenostoma stations is in nearly every case higher than the
sun temperatures, while in the other stations the reverse is true.
The reason in the former case is not clear.

58 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

The maxima during the dry season would often be considerably
higher than those recorded here. Even in winter the day tem-
peratures are frequently quite high. Frosts are rare and slight.
During the winter of 1913-14 no atmometers were broken from this
cause — and these instruments are very sensitive to that danger.

Table 10. — Atmospheric temperature.

July 4

, 1916.

Aug. 16, 1917.

Sun.

Shade.

Sun.

Shade.

Station No. 1 . . .

27.5

30

22.8

24.4

2. . .

25

28.5

22.5

23.9

3. ..

27.7

25.5

21.7

20.3

4. ..

30

29.5

23.6

23.4

5. . .

29.5

30

23.9

25.0

6. . .

28

25.7

22.8

21.1

7. . .

27

26

20.8

18.9

8. ..

28

25.6

20

19.4

9. . .

26.3

27.5

21.7

20.6

10. . .

26.6

26

20.6

21.7

Summary :

Adenostoma. . .

27.5

28.6

22.5

23.2

Arctostaphylos .

27.8

25.6

22.2

20.7

Forest

27.5

25.8

20.4

19.1

Wind.

I have no data for Jasper Ridge, but a statement as to the sur-
rounding region as a whole will be of value. Throughout the Santa
Cruz Peninsula and northern Santa Clara Valley, except where
topography produces local differences, the prevailing direction of
wind is northwest. During the dry season this is almost universally
true. During the wet season the direction is much more variable.
Northwest winds are still the commonest, but the severe storms
that bring the rains are usually accompanied by gales from the
southwest. The north-facing slopes, therefore, receive most of the
dry summer winds, while the south-facing slopes receive the full
force of the rain-bearing winter gales. That the former may be
very severe in their effects I have shown in another paper (23,
p. 187). The actual vegetation cover of the north- and south-
facing slopes is evidence that in general the prevailing winds are
not important as locally controlling factors.

The Atmospheric Moisture.

Rainfall, cloud, and jog. — These topics have already been treated
(p. 31). They are of interest in consideration of the locality as
a unit, but have little influence upon the distribution of the plant
communities as determined by slope exposure.

Relative humidity.— The effects of relative humidity upon veg-

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 59

etation are included when the evaporating power of the air is
measured, as are also in part the effects of temperature, wind, and
light. The two series of observations shown in table 11 are of
interest, however, in illustrating the values of this simple factor in
the various stations. The readings were made with a cog psy-
chrometer, breast-high, except where dense chaparral made a
lower level necessary.

Table 11. — Relative humidity, percentages.

July 4

, 1916.

Aug. 16, 1917.

Sun.

Shade.

Sun.

Shade.

Station No. 1 . . .

48

43

53

51

2. . .

57

46

59

54

3...

48

49

60

60

4. ..

44

42

56

52

5...

42

39

54

48

6...

45

49

57

59

7...

49

48

61

66

8. . .

45

47

62

66

9. . .

46

42

60

63

10. ..

54

48

63

60

Summary :

Adenostoma. . .

48.5

43.3

57.5

54.7

Arctostaphylos .

46.5

49

58.5

59.5

Forest

47

47.5

61.5

66

The percentages given in table 11 do not approach to what we
are accustomed to think of as xerophytic values. It may be that
they do not represent average dry-season conditions, though there
is ample reason in the proximity of the bay and the ocean for
relatively high humidity. Other chaparral localities would cer-
tainly show a much lower average. The sun humidities in the
Adenostoma habitats are higher than those in the shade; in the other
stations the order is reversed.

Evaporation. — Observations were made by means of the Livingston
cylindrical porous-cup atmometer, standardized at the beginning and
end of their use by the Plant World Company. For dry-season
study instruments were set out in all the stations except No. 10,
two per station, one at the level of the ground, and one near the top
of the vegetation in the region of most abundant foliage. The
instruments at the ground-level were set in small excavations, so
that the tops of the porous cups were about 2 dm. above the surface.
Those at the upper levels were set upon posts of proper height,
ranging from 0.7 meter in station 1 to 2.5 meters in station 6. In
the forest stations, No. 7 and 8, it was necessary to construct wire
trolleys running to the tops of trees, so that the instruments could
be lowered for measurement.

60 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

Considerable trouble was encountered by reason of the depreda-
tions of thirsty animals, and several of the surface cups were de-
stroyed, so that the lower series is incomplete. The dry-season
study was carried on from June 8 to October 17, 1913, and readings
were taken at weekly intervals, with a few irregularities. For the
investigation of rainy-season conditions, stations 3, 4, and 7 were

<* <t <* <* c

A VCD d3d Sy3-L3tMUN3D 31803 'NO UVd0dVA3

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 61

selected, representing the Adenostojna, Arctostaphylos, and forest
habitats, and rain-correcting atmometers were substituted for the
ordinary kind. The surface instruments were surrounded with
chicken wire, which proved an adequate defense. Readings were
taken weekly until the close of the soil-moisture observations on
January 16, 1914, and at intervals of approximately four weeks

AVERAGE DAILY EVAPORATION IN CUBIC CENTIMETERS

Dry season Wet season

131 days, June 8 to October 17, 1913 114- days, November 10,

' 1913 to March 4-, 1914

Sta.l. 5outh slope
Adenostoma 30.4

Sta. 2. Summit
Adenostoma

Sta. 3. North slope
Arctostaphylos

Sta.4. South slope
Adenostoma

Sta 5 South slope
Adenostoma

Sta 6. North slope
Arcrostapriylos

Sta 7. North slope
Forest

Sta. 8. North slope
Forest 25.3

Sta.9. South slope
Adenostoma 27.7

Summary, dry season
All Adenostoma

All Arctostaphylos

28.2

Maximum daily evaporation. week ending September 19
Sta.l

49.6

Sta. 2

51.4

Sta. 3

48

Sta. 4

49.5

Sta. S

48.2

Sta. 6

47.6

Sta. 7 38.6

Sta 8

47.4

Sta. 9 44.2

Fig. 15. — Average daily evaporation in cubic centimeters in
stations 1 to 9 at Jasper Ridge, for the dry season
of 1913 and the succeeding wet season; also the
maximum daily rate (week ending Sept. 19).

thereafter until April 17, when the series was closed. We thus
have records of one complete dry season, following a winter of
deficient rainfall, and of one complete wet season — an unusually
rainy one — fair samples of extremes in both directions. The dry-
season series is complete for nine stations at the upper level and
incomplete for the lower; the wet-season series is practically com-
plete for the representative stations at both levels.

The results have been summarized in three figures. In figure 14
the upper-level series is used, and the water-losses for three rep-

62 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

resentative stations (Nos. 2, 3, and 7) are given. These, as pre-
sented, denote the average daily evaporation in cubic centimeters
for each observation period, reduced to the official standard main-
tained by the Laboratory of Plant Physiology of Johns Hopkins
University. The conclusions may be very simply stated. The high
rate of evaporation for the dry season and the low for the rainy
season are immediately evident. A rainless period in December
caused a sharp rise. The relative positions of the three lines are
maintained throughout the period of observation almost without
change, the Adenostoma habitats having the highest rates and the
forest the lowest. The absolute maximum attained was 51.4 c. c.
in station 2, Adenostoma chaparral, for the week ending September
19. This period included some of the hottest days recorded for the
region. The absolute minimum was attained on January 9, 3 c. c.
daily evaporation for the period of 6 days preceding.

Evaporation

R +* J To P °* Vegetation CUnity)
Katl ° V Surface of Ground
Dry Season Wet Season
49 days 114 days
June 8 to July 27 7 Nov. 10, 1913, to
1913 March 4, 1914
Sta. 2 Adenostoma
I J

".92

Sta.3
1

.73

Sta. 7
1

.82

.53

Arctostaphylos

Forest

J

.69

Fig. 16. — Ratio between evaporation at top of vegetation and at surface of ground,
in three representative stations at Jasper Ridge.

In figure 15 the average daily evaporation for each station is
given, for both wet and dry seasons, readings from the upper series
of instruments being used. During the dry season the Adenostoma
stations have regularly the highest rates. Station 2, located on
the topmost point of Jasper Ridge, exposed to winds from every
direction, naturally has the highest of all. Stations 1, 4, and 5,
on gentle south-facing slopes, come next, and station 9, on the
steep south-facing slope of an abrupt ravine, is last, falling slightly
below the Arctostaphylos stations. The latter (stations 3 and 6)
are closely alike, and fall between the Adenostoma and the forest
stations, but nearer the former. The forest stations are decidedly
below the others. Evaporation during the wet season averages
about 30 per cent of the dry-season rate in the stations studied.
The figure also shows rates of evaporation during the most extreme
conditions of the year. The figure for station 8 seems abnormal.

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 63

Figure 16 presents the ratio between evaporation at the top of
the vegetation and at the surface of the ground. Stations 2, 3, and
7 were selected, since they were continued through both wet and
dry seasons, and the average daily losses for the period are used
in computing the ratios. During the dry season the rates at the
two levels did not differ very greatly, while during the wet season
the differences were considerable, especially in the chaparral stations.
This suggests that during the latter period there is a surface stratum
of moist air due to evaporation from the saturated soil, and that
the contrast between it and the upper air layers is greater in the
chaparral stations than in the forest because of freer circulation.

DISCUSSION AND CORRELATIONS.

It is very easy and pleasant to describe the vegetation of a locality,
to mark it out into communities, and to name the dominant and
secondary species thereof, and even to present quadrats of the same.
It is not so easy completely to analyze the habitat. With our
present methods of investigation certain factors elude us almost
altogether, and much of the data, even when obtained, is difficult to
utilize. It is most difficult satisfactorily to link vegetation and the
measured factors of the habitat in the relation of cause and effect.
The present chapter has so far been devoted to description of veg-
etation and analysis of habitat. Some correlations of the two will
now be attempted.

THE BROAD-SCLEROPHYLLS AND THEIR HABITAT.

After the analytical treatment that has preceded, it is proper
first to synthesize the various elements into a whole, to characterize
the broad-sclerophyll habitat as a unit, so far as we fairly may from
the data in hand. Afterward it will be in order to consider habitat
differences and community differences and their correlation.

To describe the broad-sclerophyll habitat in the large is merely
to describe the climate of California west of the Sierras and its
effects upon the soil factors. Emphasis must be placed upon the
complete seasonal march of the factors, since vegetative activity of
some sort and degree may here take place the whole year through.

Of direct and fundamental importance is soil-moisture. Its
great abundance during the winter months makes ample provision
for growth during that time. The supply is made more completely
available by the low air-temperature and low evaporation-rate
during the period. The cessation of the rains at some time during
the spring is followed by gradual depletion of the water-content,
due to failure of the renewal supply, gravity drainage, evaporation
from the soil, and absorption by the vegetation itself. These
partial causes will vary in importance in various places. In a

64 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

region of relatively heavy rainfall, the dry season is likely to be
shorter and the supply will dwindle less rapidly. A soil with a
high water-retaining capacity will lose less through gravity. If the
evaporation-rate is high, more will be lost thereby, and forest
vegetation will extract a far greater quantity than will scrub. But
whatever the difference in rate due to these separate elements, the
depletion goes on, and the water-content approaches the wilting
coefficient. In some places at least, and in some years, it actually
drops below for a considerable period. The Jasper Ridge observa-
tions demonstrate this; moreover, the general conditions of moisture
here are much less severe than in other regions where broad-
sclerophylls are even more thoroughly in control. When the autumn
rains finally arrive there is still, in some soils at least, a period of
several weeks before the water forces its way into the deeper layers
of the air-dry soil. The importance of this delay is lessened some-
what by the fact that much of the root system is contained within
the first half-meter of depth. There is probably but small delay
in the production of new absorption apparatus from the old dormant
roots, once the moisture touches them.

As the store of soil-moisture decreases, the evaporation-rate rises
and water-loss from the leaves must increase. Thus the water-
balance is affected unfavorably at both ends at the same time, and
the late summer and early autumn, therefore, constitute the time
of greatest danger. The deeply penetrating roots and the effectively
guarded leaves are the answer.

Air-temperature and, in response, soil-temperature, at a minimum
during the months of abundant moisture, rise as water-content
decreases. This combination is altogether unfavorable for veg-
etative activity. When moisture is abundant, the low soil-tem-
perature makes absorption difficult and root-growth slow, and the
low air-temperature is unfavorable to photosynthesis and the other
growth activities. When the soil-temperature is favorable for
root-growth and absorption, the water-content is scanty or neg-
ligible. When air-temperature is favorable to rapid photosynthesis,
the water supply necessary to that process and to vigorous growth
is insufficient. High temperature is now itself a danger, in that it
raises the evaporation-rate and thus increases the depletion of the
scanty moisture supply, both in plant and soil.

The seasonal development of the broad-sclerophylls is accurately
adjusted to the peculiar seasonal march of the factors just outlined.
With the beginning of the heavy rains (usually December, sometimes
earlier or later), growth starts in most of the species. The history
in detail presents many variations. For instance, Arctostaphylos
flowers immediately from buds already formed, then proceeds to
put forth new leaf-shoots, which terminate with next year's flower-

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 65

buds. When these are fully formed (about May), growth for the
year practically ceases. In Adenostoma the procedure is exactly the
reverse. This species begins its activity with the formation of its
vegetative shoots, and its flowers, borne on these, open in June and
close the period of active growth. All agree in making the fullest
utilization of the relatively short period when both moisture and
temperature conditions come nearest to combining for the general
good, i. e., in the spring months, when soil-moisture is still abundant
and air and soil temperatures are on the increase. April, in the
Palo Alto region, is the month of greatest vegetative activity among
the broad-sclerophylls and other types of plants as well. An
additional advantage which many of the former show at this time
is the possession of an increased photosynthetic apparatus, since
the last year's Leaves have not yet fallen and the new ones have
reached or are approaching maturity. It is doubtless at this period
that the superficial roots described on page 91 have their greatest
extension. A glance at figure 8 will reveal the interesting fact that it
was in April that the soil-moisture suffered the most rapid depletion —
a fact that may reasonably be referred to the abundant use of water
by the vegetation at this time.

Soon after the period of maximum activity, the decreasing water-
supply makes a limitation of growth inevitable and necessary, and
it is surely more than a coincidence that a number of the species
drop their oldest leaves rather regularly during the month of June.
This phenomenon is sometimes so constant and definite that the
chaparral may take on a distinctly yellowish hue for a few days,
resuming its normal tone after the old leaves have dropped off. The
same is strikingly true of the broad-sclerophyll tree Arbutus, which
passes through a period in June when dead leaves are its most
conspicuous feature. It is of interest to note further that Msculus
californica, not an evergreen, has a corresponding habit, dropping
its leaves gradually during the summer, apparently as a response to
the decreasing water-supply, until it is often nearly leafless before
autumn. After the period of leaf fall the broad-sclerophylls enter
a state bordering upon dormancy, growing very little, merely main-
taining themselves through the scanty supply of water absorbed
by the few roots that penetrate the deeper soil layers.

That water lack is the reason for the stoppage of growth activity
is shown by the behavior of stump sprouts. If the aerial portions
of a shrub are removed during the dry season, even in the most
severe part of it, sprouts are immediately put forth which grow
with amazing rapidity. On Mount Tamalpais, three weeks after
a severe fire, I found sprouts of scrub oak 45 cm. high. Visiting the
scene of the Ojai Valley fire of June 1917, six weeks after the event,
sprouts of 90 cm. were found. In such cases the water-supply of

66 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

a whole clump of shoots is suddenly concentrated upon a few small
sprouts. Moisture and temperature are both favorable to the
utmost, and strikingly rapid growth is the result.

The meaning of the evergreen sclerophyll is plain enough when the
seasonal factor-march is thus considered. It is a compromise which
enables the plant to meet the lack of agreement in time between

!S r~

R

CD

l_

o5*
r cliu
-ij ro c
to x : o

LOU-

o :
a. : ;
<0 • •
> : :

hi : :

~m

p to ji

~£ Q-<u
™ to (-

gyu.

e : i

U ' ■
-P i !

%\\
o

— 4->

•-+J

to rain

<- Q-QJ
D CO u

~-

,;\

i

i

T-

m<

/

2?-)ff> —

3 £

<f^" —

10

Tf

_

* : •>

'»"

73
®

<-> ^ —

£-4

<2

>>

*

^

in —

-

o

Z N

'

O ,,

.

-

-~

i .■•

1

\

I

c--

■a

S

.-

1

I

r.

o

"7

,'

/

/

.a

o

1

•-.

*■

""1

..

\

c

-

..

■ <\

,,<

'..

•a d

X) -

f

"

'

o-2

N

,,

-

'"

r*

^

...

—

)

la

a. -

• v

■ii

.-

.

">n -

-*-
1

,.

/

1>

-

j

,,

S N

6"

\

v-

1

_ *

^ 5

e3 co

■S s

< '

<*"

1

t ,4'

- ~

..-

- J

s

"T

•-.

tl

n -

'"'"

—

\

> a

-"

\

. .■•

I

—

I

/

O i

-i -

*^-

*.-.'

••\

!-

1

\

a S _

B

^

«»-

\

"

•*"

S|

z "J.

/

.

3- "

L --

>

■

-

^

\-

So

->9 -

,.

<■'

-'

V

/

•-a 59

"

' —

—

1—

5»

r^

...

•-

N

s

y-

*= o

< M "

.,

\

'

■--

V

S3

10 -

s,

♦V

?.s

/

•»" t.

K m "

*

/

< "

."£ t^

--

-..

\

2 o

-

■»-

'S

K w

"o -3

<5 -

5

E

-C ^

e3

■t

Hi

>>

Id ~

s

p

u

-•

g

■*

s

N

3

. rj

I

*<

^*-

CO

1

* fJ

Js

'•

S^ 2

-FP

J l ^

1/

o s

t -

9

o

U3J

3 IX'

in

.3

•x

o

All

d

33

a

3 3

i

in ui
pj t\

o
in

m

tu

•*

fl

4
3 \

3c

o

to

S

01

a

ei

a:
to

aa

f
n

X.

Ch

r>

3 1^
10

rd

IS

o
rj

IX

in

01.
in

CO

w

N3
•*

LN

N
O

3D
m

f

>IG

3i

'J c
N PJ

no

- c

jnx

— c

CD

'NO
01

va3

Ol

IX
CD

dl'
8}

4 t>

vac
s it

>I3X

o

dV/
in

yi

CD

i3

iC 4
o ft

t-j
-
—

o

2

o M
o

XH9I3AA ASJO JO 1KJ3D «3d '3ynxSIOVN "IIOS

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 67

moisture and temperature optima. The prominent cuticle and the
other features that are effective in limiting water-loss enable the
plant to exist through the critical period of late summer and autumn.
The evergreen habit permits it to make use of the spring period,
when growth conditions average best — a time when an ordinary
summer-green plant would be putting all its energies into the build-
ing of a new photosynthetic apparatus — as well as the rather frequent
periods during the winter months when the temperature is high
enough for effective photosynthesis. This explanation has been
advanced by Guttenberg (36) for the broad sclerophylls of the
Mediterranean region, and by Schimper (80) for the type in general.
It seems entirely adequate for the Californian species. It might
be added, however, that possibly in JEsculus californica we have
a still more effective and economical way — the quick production
of thin, deciduous leaves for use during the favorable period and
their gradual elimination as unfavorable moisture conditions render
them dangerous rather than useful.

In figure 17 I have combined the principal factors of the broad-
sclerophyll habitat in a way that gives a vivid impression of the
whole — rainy and dry season, decreasing summer water-content,
with increasing temperatures and evaporation.

THE HABITATS OF THE TWO CLIMAXES CONTRASTED.

The topic that naturally follows is comparison of the respective
habitats of the two broad-sclerophyll climaxes, forest and chaparral.
Both communities are typically represented on Jasper Ridge,
occupying opposite sides of the same ravines, and two subcommunities
of the chaparral as well. Care must be exercised in extending the
conclusions here drawn to the two climaxes in general, for it is of
course true that the factor values noted upon the slopes of a single
hill do not in detail perfectly represent the averages that characterize
great regions. It may be said, moreover, that given freedom from
disturbing agencies, the more xerophytic community, through the
reactive influence of the vegetation, might continue its mesotropic
development until forest replaced chaparral. Certain determination
of this point in any single given case is very difficult. Detailed
examination of the tension areas at Jasper Ridge failed to produce
satisfactory evidence, either positive or negative. For the present
purpose, the question may be settled with sufficient certainty by
treatment in line with the discussion upon page 74.

Theoretically, there must be a region intermediate between those
clearly dominated by forest and by chaparral where the control is
in suspense, and here any decided local variation in habitat toward
xerophytism or mesophytism, as would result from opposite hill or
ravine slopes, would bring about permanent control of a limited

68 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

area by one or the other climax. All the evidence points toward the
southern Coast Ranges as a region of this nature, and it therefore
seems safe to make use of the data obtained at Jasper Ridge in
illustrating the differences between the larger units. Further, the
differences here are likely to be close to the minimum in degree
which are able to bring about the differentiation of climaxes, and
are of special interest on that account. Such a comparative evaluation
must of necessity include the reactional effects of vegetation upon
habitat, accumulated through a long period of development. This
is as it should be, since in the interplay of primary and reactive
factors we find the "continuing causes" of the vegetation.

Summarizing and commenting upon the differences between
forest and chaparral habitats, we attain the following results:

As to soil, humus in the chaparral is very scanty, but in the
forest is abundant — nearly 2 per cent by weight in the surface
layer and considerable to the depth of 1 meter. In water-content
there is large difference during the rainy season, the forest having
the greater amount. At this time the surface layers are most
important, since the major part of the absorbing roots is contained
therein. It is here, too, that the water-content differences mainly
show themselves, being practically negligible at the depth of 1
meter. As the dry season advances, water-content values in both
communities and at all depths converge, and at its culmination they
are all very close together, and the correspondence is rendered still
more striking by comparison with the wilting coefficient in each
case. In brief, there is notable difference in the actual amount of
water available, but at the critical period conditions are about
equally severe in both communities. In water-retaining capacity
the only noteworthy feature is the relatively high value in the
surface soil of the forest community, due to humus. As to soil
temperature, the comparative march is the reverse of water-content;
the values are closely similar in the wet season, but widely divergent
in the dry, the chaparral being much the higher.

As to atmospheric factors, we may dismiss rainfall, cloud, fog,
and wind as immaterial to the present local problem. The light
impinging upon a leaf of the foliage canopy is much greater in
chaparral than in forest, because of the fewer obstacles to its trans-
mission and the reflection and diffusion from the light-colored soil-
surface. The intensity in the shade is considerably less beneath the
forest canopy, both absolutely and proportionally. The fact that
the shade intensity beneath Arctostaphylos is practically the same as
in the forest indicates that the leaf character is determinative — the
sparse needle foliage vs. the broad leaves of the other shrubs and the
trees. Temperature and relative humidity data are unsatisfactory,
but their effects relative to the present purpose are largely included

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

69

in evaporation. The differences in this factor, though not strikingly
great, are constant throughout the year, the Adenostoma chaparral
being the highest and the Arctostaphylos chaparral intermediate.
This conclusion is drawn from the values obtained at the top of the
vegetation. The rate at the surface of the ground does not show
differences of import to the problem in hand.

It may be objected that the evaporation values used here are
affected to an unknown degree by the presence of abundant vegeta-
tion. In order to discover something of the actual comparative

SLOPE PROFILES

N.

rSta.I

S.
Sta.lly

N.
\,Sta.lV

S.
Sta.lII

^y^.

-M—^**

^pV.

^i?

50

ioo Meters

JULY
7 16 23 29

AUG.
12 19 26

JULY
7 16 23 29

AUG.
7 12 19 26

\

?

/

\

1

\,

\

? .

v\

1

/

\

L \

/ I

\

i

/

/

\

\

\

//

\

\

\

\

/

>

V

1

1

\

/

\

\

/

^

^

/

1

f

\

/

\

/

1

Sta.I (S-facing)
Sta.lI(N-facing)

Sta. IV (S- facing)
Sta.lII (N-facing)

Fig. 18. — Evaporation during the dry season of 1915 on opposite
north- and south-facing bare slopes. Foothills of the
Santa Cruz Mountains, near Palo Alto.

values upon opposing north and south slopes, uncomplicated by
reaction, a short study was made in the hills near Stanford Uni-
versity in the summer of 1915, an account of which is inserted here.
Evaporation was measured upon the contrasting north- and south-
facing slopes of two ravines, distant about 5 km. from Jasper Ridge.
The profiles of these are shown in figure 18. Stations 1 and 2
were 15 meters, and stations 3 and 4 were 18 and 15 meters, re-
spectively, above the valley floors. The slopes were bare of shrubs
and trees, the vegetation being largely wild oats, which, with the
other herbaceous plants, were dead and brown at the time of the

70 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

study. The stalks were cleared from the vicinity of the instruments
and the effects of vegetation were thus eliminated. The atmometers
were installed July 1, and were visited at approximately weekly
intervals until August 26. Unfortunately, three readings at station
1 were lost through interference by cattle, which did not respect
the chicken-wire defenses. Except for one reading in stations 3
and 4, the south-facing slopes show the higher evaporation-rate.
The reversal on July 29 is hard to understand, but may possibly be
due to experimental error. Omitting the figures in station 2, which
have no corresponding south-facing slope values, we find the following
average evaporation-rates :

For the first pair, south-facing : north-facing : : 100 : 85.

For the second pair, south-facing : north-facing : : 100 : 79.

For the 13 complete readings afforded by the two, south-facing : north-facing : : 100:82

There is, then, a distinct difference in evaporation-rate between
opposing north and south slopes, when uncomplicated by vegetation.
In these open ravines with comparatively gentle slopes the average
rates were as 4 to 5. If they had been as steep-sided as certain ravines
at Jasper Ridge, the contrast would doubtless have been greater.

We have here sufficient data to explain the greater size and
luxuriance of the plants growing upon the north-facing slopes, and
the absence of the more mesophytic species from the south-facing
exposures. Many factors are indirectly involved, but fundamentally
the problem is one of water-balance. During the height of the
rainy season there is in both habitats an abundance of soil-moisture
and a minimum of evaporation. But the air and soil temperatures
are low, and there is consequently little growth. As the temperature
rises, the rate of growth increases accordingly, but at the same time
the rains are diminishing in frequency and amount, finally ceasing
entirely, and the evaporation-rate is going up. It has already been
stated that the period when most of the growth occurs is in the
spring months, culminating in April. The amount of actual growth
during this time, other things being equal, will depend upon the
ratio of water-supply to water-loss, and plainly this water-balance
is more favorable upon the north-facing slope inhabited by forest.
The size of the full-grown plant of a given species depends largely
upon the turgor of the growing cells, and a healthy condition of
turgor is conditioned upon a water-supply that will more than
compensate for the loss by evaporation. If the supply is so scant
or the loss so severe that the water-balance is barely maintained at
equality, the plant and its organs will be small. If either member of
the system varies in such a way that the supply is well in excess of
the loss, rapid growth will result. The ever useful example of the
stump sprout illustrates this point. The water-supply which
formerly fed a full-sized shrub is concentrated upon a few small

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 71

shoots, which grow for a time with surprising rapidity (see p. 86),
although no decrease has occurred in the rate of loss. The size
and luxuriance of the plants are therefore greater upon north-facing
slopes, where soil-moisture is more abundant and evaporation less
rapid during the growing-period.

To account for the absence of the more mesophytic species from
the south-facing slopes, we must appeal to the dry season. It has
been shown that the soil-moisture contents of the two slopes, con-
sidered in relation to the wilting coefficient, are about equally unfavor-
able during the critical period. This is not true, however, of the
evaporation-rate, which is considerably higher in the chaparral
community. A seedling of a mesophytic forest species, which might
conceivably obtain a start during the spring months, though not
making much increase in size, in the chaparral habitat, would be
very likely to succumb during the ensuing critical period, when the
high evaporation-rate had made the water-balance still less favorable.

Since the plants of the forest, because of their size and abundance,
use more water than the chaparral species as food material and lose
more through their enormously great expanse of foliage, more is
withdrawn by them from the soil. The much more rapid depletion
of the water-content during the spring months, resulting from this,
is perfectly evident in figure 8. Thus the very luxuriance of the forest
vegetation is itself the prime cause of the reduction of water-content
practically to the low level of the chaparral.

Our conclusion, then, extended to the larger vegetation units,
is that the fundamental distinguishing difference between the two
broad-sclerophyll climaxes — their continuing cause, so to speak — is in
the water-balance and its variations, whatever the indirect factors
influencing it; that its importance is equally divided between wet
and dry seasons, the greater excess of supply over loss in the forest
during the growing-season explaining the size and luxuriance of the
plants living there, and the higher evaporation-rate in the chaparral
during the dry season, with equally severe soil-moisture conditions,
accounting for the absence of mesophytic species in that habitat.

72 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

V. DEVELOPMENT.

THE CLIMAX REGIONS OF CALIFORNIA.

In a region of slight relief and uniform climate, such as the Missis-
sippi Basin or the Great Plains, a single climax extends for hundreds
or even thousands of miles. In California the case is very different.
The extremes of latitude and altitude, the proximity of the ocean,
and the barrier influence of high mountains upon moisture-bearing
winds all operate to bring about great and abrupt climatic differences,
and therefore great vegetational complexity. Instead of a single
climax community there are a half dozen, each, though limited in
area, comparable to such a community as the Great Plains grassland
or the eastern deciduous forest. Successional investigation has not
proceeded far enough to enable us to define these climaxes with
accuracy. The following outline is tentative, and is given merely
that we may properly place the communities with which we are
dealing.

The conifer-forest climaxes are three in number, and the first two
have important contact relations with the broad-sclerophyll com-
munities. The Pacific Conifer Climax Formation, which includes
the forests of the coastal region from California to Alaska, is repre-
sented in northwestern California by two associations: (1) The
Sequoia sempervirens association ranges from the northern boundary
of the State, in the immediate vicinity of the coast, southward to
San Luis Obispo County. South of San Francisco Bay its continuity
is much broken. (2) The Pseudotsuga association, which has its
center in the Puget Sound region, is poorly represented in California,
occurring in the interior of the north Coast Ranges. It is confused
with other communities, especially the broad-sclerophyll forest,
so that the portion included within the State should probably be
considered as transitional.

The Montane Conifer Climax Formation covers the middle altitudes
of the interior mountain region of western America, and the character
tree everywhere is Pinus ponderosa. In California it is represented
by the Pinus ponder osa-lambertiana association, which occurs in the
middle altitudes of the west slope of the Sierras and on the higher
Coast Ranges and the mountains of southern California.

The Subalpine Conifer Climax Formation is practically coextensive
with the preceding in general range, but lies above it altitudinally.
I will not apply a designation to the California subdivision, as it is
rather complex, and no successional work has been done within it.
It occurs in the higher Sierras and there are isolated outliers else-
where. The character trees are Abies magnifica and Pinus monticola
in the lower part, and Pinus albicaulis and Tsuga mertensiana near
timberline.

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 73

The broad-sclerophyll climaxes, including the broad-sclerophyll
forest and the climax chaparral, will shortly be treated in detail.
The alpine meadow climax belongs to the highest summits of the
Sierras. The desert climax occurs east of the Sierras and the moun-
tains of southern California. The basin sagebrush climax enters
the eastern part of the State to some extent, and the coastal sagebrush
is locally important in the interior valleys of the south. It is possible
that the Great Valley or a part of it is climax grassland. The
coastal sagebrush and the grassland are in extended contact with the
climax chaparral.

THE BROAD-SCLEROPHYLL CLIMAX FOREST.

Discussion of the climax nature of the broad-sclerophyll forest
involves considerable difficulty. Its distribution is patchy, i. e.,
no extensive area is completely dominated by it. In parts of the
region elements of other vegetation types are mingled with it, giving
the appearance of a transitional type. Its natural distribution has
plainly been restricted somewhat, especially by fire.

The range of the broad-sclerophyll forest is practically coextensive
with that of the climax chaparral, and over most of the region there
is sharp alternation between the two. The forest is least important
southward. In the Cuyamaca Mountains the chaparral covers all
slopes of the lower altitudes, the forest being confined to the steepest
ravines and similar situations where the moisture conditions are
unusually favorable. As we go northward, or upward in the moun-
tains, we find the forest increasing in importance. In the central
Coast Ranges it is the regular covering of the steeper north-facing
slopes, the chaparral being confined to south slopes and summits.
In the San Francisco Bay region the two are roughly equal in impor-
tance, and still farther northward the forest continues to gain.
Here, however, a new complication arises. Pseudotsuga mucronata
is present in the most mesophytic situations and rapidly increases
in importance northward. In some places it forms pure growth, but
oftener occurs in mixture with the broad-sclerophyll trees, the chapar-
ral still holding the south slopes, but steadily decreasing in impor-
tance. Finally, we encounter a clear dominance of Pseudotsuga,
with the broad-sclerophyll trees forming an understory. Ascending
one of the southern mountain ranges, we meet with a similar tran-
sition, the broad-sclerophyll forest gaining over the chaparral and
finally merging with a forest of Pseudotsuga, the species of the last
being different (P. macrocarpa) , but to a degree ecologically equiva-
lent to the northern one. In short, the broad-sclerophyll forest is
plainly unable to control in the south; it increases in relative impor-
tance northward and upward, but just as it begins to show decided
signs of dominance over the chaparral, it comes into competition
with a coniferous element, to which it soon becomes subsidiary.

74 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

Before proceeding to interpretation, certain definitions are neces-
sary. The climax is here considered to be the most mesophytic
community which a region as a whole, under present climatic con-
ditions, is able to support. Regions nearby have different climaxes,
more mesophytic or less so, according to whether the climates are
more or less humid, and the climax of the first region might be
succeeded by one or another of these, should the climate change in
the appropriate direction. The neighboring climaxes are thus
potential climaxes for the region under discussion (Clements, 21,
p. 108). A potential climax is a postclimax if it is more mesophytic
than the existing one, and a preclimax if less so (1. c, p. 109). Fre-
quently a potential climax is actually present in a region, in situations
of unusually favorable or unfavorable moisture conditions. An
excellent example of a postclimax is the forest which follows the
river valleys far westward into the Great Plains region.

In the light of these definitions it is evident that the broad-sclero-
phyll forest is postclimax in Southern California. The chaparral
is plainly in control, although it is certain that its dominance has
been extended somewhat by fire. In central California, where the
two are rather evenly balanced, the question is difficult, but the
tendency of the forest northward to increase its dominance is plain.
In such a transitional region, where the actual control is uncertain,
the spatial relations of the competing communities depend upon the
nature of the topography. In a level country the change is usually
seen in a gradual mingling of the species of the two impinging regions.
In a broken country, where steep slopes with various exposures are
the rule, the situation is otherwise. The differences in atmospheric
factors resulting from contrasting slope exposures are in such a region
no greater than elsewhere; but because there is here a zone in which
the general conditions permit both groups of species to flourish,
the relatively slight differentiation of habitat due to the factor of
slope exposure exerts a maximum apparent effect upon the vegetation,
a sifting of the species into two groups. We therefore find the
two communities sharply set off from one another, each in permanent
control of its area, with no successional relations between them.
Each is the climax within its own little sphere of influence. Such is
the relation of broad-sclerophyll forest and chaparral in the central
Coast Ranges, the forest permanently controlling the north slopes,
the chaparral the south. Passing into the north Coast Ranges,
we would expect to find the broad-sclerophyll forest finally attaining
a true climactic status, and there is an undoubted tendency in this
direction. The effect is spoiled by the competition of Pseudotsuga,
the addition of which produces a mixed broad-sclerophyll-conifer
community, which is apparently the climax of the region in which it
occurs. Farther north, the broad-sclerophyll element loses all

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 75

claim to dominance, becoming a subordinate part of the conifer
forest, and possibly a successional stage in its development.

In way of summary, we may say that although no extensive region
is completely dominated by broad-sclerophyll forest, nevertheless it
must be considered a climax, since it represents a degree of meso-
phytism between that of the chaparral and the conifer forest, and
because it shows plainly a tendency to gain dominance over the
chaparral, attainment of which is hindered only by the competition
of a still more mesophytic type.

THE CLIMAX CHAPARRAL.
EVIDENCE OF ITS CLIMAX CHARACTER.

The evidence of the climax nature of that type of chaparral that
I have designated as such is presumptive and direct. Under the
first head four points may be made.

Dominance. — The very fact that the chaparral is by far the most
widespread of all the communities within its range (this being espe-
cially true in the region of its center of distribution) in itself creates
the presumption that the chaparral is the climax. Bearing in mind
that there are cases in which an important and widespread community
is not the climax of its region, for instance, the Pinus murrayana
forest of the central Rockies, whose temporary dominance is due to
fire, we can not accept this evidence without confirmation.

Stability. — In most areas the chaparral has obviously been in
control for a long period of time — so long that there is no evidence
existing of previous vegetation of other type. This is not universal,
and in any case is not entirely conclusive.

Occurrence on diverse sites as to soil and topography. — The fact
that a community is not restricted to a single topographic situation
or soil type, but occurs on many, is good evidence of climax nature.
This is true of the chaparral in the region where other evidence
indicates its climactic character, i. e., its center of distribution.

Adjustment to climate. — In every region there are many plants
which are plainly in adjustment with some very special habitat.
Such are aquatic plants, halophytes, and species of rock crevices.
All of these occur in the chaparral region. But the species of the
chaparral itself possess a constant and much specialized character
which is plainly a response to a very particular type of climate.
The correspondence in area covered by climate and vegetation type
has already been shown. If further evidence be desired, it may
be derived from the ecologically equivalent "macchie" of the
Mediterranean region and its relation to a climate which is almost
identical with that of California. It is difficult to escape the con-
clusion that a vegetation type which is so perfectly adjusted to the
climate of its region must be the ultimate or climax type, in that
area at least where it reaches its most perfect development.

76 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

The direct evidence for the climax nature of the chaparral is
derived from a study of the successions themselves, in which we
find that all that have been investigated in a rather widely extended
exploration culminate in the establishment of this community, and
that over much of the region there is no evidence that any other
community is superseding it. Detailed consideration of the succes-
sions is reserved for a separate section.

EXTENT OF THE CHAPARRAL AS A TRUE CLIMAX.

Because of widespread disturbance by various agents of destruc-
tion, it is impossible to set definite and certain limits to the region
of true chaparral dominance. Fire occurring in the transition zone
between two formations seriously alters the normal relations due to
climatic control. Another source of difficulty lies in the fact that
much of the original vegetation of the economically more useful land
has been totally destroyed by the invading white man, so that it is
necessary to reconstruct the original picture from fragments. Certain
areas, however, may be fixed upon as chaparral climax which will
not be questioned.

The foothills and low mountains of southern California and certain
parts of the valleys and mesas are the most certain, and it is probable
that exploration of northern Lower California would add a large
extent of country to the range of the chaparral climax. In these
places Adenostoma is usually the dominant species, or if cultivation
has largely destroyed the native vegetation, remnants show it to
have been so at a former time. Such is the case in portions of the
Los Angeles and San Bernardino Valleys and on the broad mesas
east of San Diego. As we ascend the higher mountain ranges of
southern California, the evidences of climactic character in the
chaparral become less and less convincing. Species which belong
to the conifer-forest chaparral appear, and there is more and more
alternation with patchy areas of broad-sclerophyll and of conifer
forest, in which Pseudotsuga macrocarpa is the first to attain impor-
tance. In this region there has plainly been an increase of the chapar-
ral areas at the expense of the forest, due to the fact that repeated
fires in a transition zone favor the extension of the more xerophytic
community. Passing northward in the Coast Ranges, we encounter
a similar transition, except that there is practically no admixture
of conifer-forest chaparral species until we are well north of San
Francisco Bay, and that the alternation is with broad-sclerophyll
tree species alone. Here, too, the forest areas show frequent evidence
of fire restriction. In the southern Sierras there is a belt along the
lower foothills where Adenostoma and other chaparral species are
of great importance, alternating, however, with much broad-sclero-
phyll forest. As we go northward, the yellow pine zone, with its
accompanying temporary chaparral, approaches more and more

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 77

closely to the Great Valley, so that roughly from the latitude of
Sacramento northward the climax chaparral is excluded from the
mountains and foothills proper, such scrub species as occur being
of the conifer-forest group. It is an interesting fact that in this
stretch of the Sierras Adenostoma is almost absent. North of Auburn
I have failed to find it in four crossings of the foothills and have
been able to discover but one record of its occurrence, noted by Dr.
Jepson northeast of Chico.

The relations between chaparral and grassland present certain
problems difficult and perhaps impossible of full solution. The
fact of the greater extension of the chaparral in the past is certain;
the magnitude of that extension is the point of difficulty. The
very general statement may be made, subject to qualifications
about to be indicated, that in central California the lower mountains
are controlled by chaparral and the plains by grasses. The character
of the transition zone between the two types is as follows:

The first hills are as a rule entirely grass-covered, though even
on these, and occasionally out upon the valley-floor, are patches
of chaparral. These show absolutely no correlation with altitude,
slope-exposure, or soil-type. Their edges are sharp and the shrubs
are uniformly developed throughout. They are obviously remnants.
Penetrating farther into the mountain mass, the chaparral patches
become more and more numerous, but are still arranged purely in
hit-and-miss fashion. Farther still, the chaparral controls the
greater area, and the grassland forms the patches, which in summer
appear like tan-colored inlays set into the green of the scrub. Finally,
the grassland disappears, leaving the chaparral in complete control.
In such a journey we do not necessarily encounter continually higher
altitudes; we are merely penetrating more and more deeply into a
mountain complex in which the ridges may all be of similar height.
In short, everywhere near the valleys and plains the hills are grassy,
while in the depths of the ranges they are covered with scrub. The
larger the extent of the mountain mass the greater is the central
area of chaparral. Conversely, a small isolated area of hills, though
of considerable altitude, may have none. This arrangement is so
nearly universal where chaparral and grassland meet that specific
examples are hardly necessary, and yet perhaps a graduated series will
be of interest.

The Marysville Buttes in the Sacramento Valley are excellent
examples of the isolated hill with practically no brush. The highest
summit is 630 meters and the slopes are everywhere covered with
the typical annual vegetation of the plains. Even here, however,
there are suggestions ol the chaparral in the presence in small numbers
of Quercus dumosa, Q. wislizeni, and Heteromeles arbutifolia (44).

Mount Diablo is also an isolated mass, but of greater extent and
height than the Marysville Buttes. Its lower flanks and the sur-

78 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

rounding hills are grassy, with patchwork of scrub, and the upper
300 meters are fairly solid chaparral, in which Adenostoma is greatly
predominant. In the Mount Hamilton Range we have an extensive
complex system made up of ridges of similar height. Penetrating
the mass at the point where Coyote Creek debouches into the
Santa Clara Valley, we encounter grassy hills, but on the first series,
and even on the footslope itself, are patchy remnants of chaparral,
mainly Adenostoma. North slopes are well forested with broad-
sclerophyll trees, and this is entirely natural, since these would
resist destruction by fire to a far greater degree than would the chapar-
ral, both by reason of their greater size and the better moisture
conditions of their habitat. Grassland is nearly continuous over
most of the slopes up to the summit of the first main ridge, where
there is a very thin forest of Pinus ponderosa, with the same meadow
vegetation under the trees. Eastward from here we look into the
heart of the mountain system, where numerous ridges are covered
with an irregular mosaic of chaparral and grassland. The Santa
Lucia Range and the mountains of San Benito County may be cited
as examples of the many systems in which one finally penetrates
to a central region of solid chaparral.

The most convincing proofs of former control of present-day
grassland by chaparral are the frequent remnants that painstaking
search brings to light. The sharply limited patches in the midst of
other vegetation, in which Adenostoma is usually most prominent,
have already been described. Experience has shown that even a
single mature specimen of this plant is almost certainly a relict and
not a fresh arrival, since when its mass control is once thoroughly
destroyed it reestablishes itself with the utmost difficulty, probably
because of special conditions necessary for successful germination,
as yet undiscovered. It is fairly safe, therefore, to assume that
wherever mature individuals of Adenostoma remain, either isolated
or in patches, that that species was formerly dominant. Fence-rows,
pieces of rocky or unused ground, ravines, etc., are the sort of places
where one looks for such evidence. Using this method, which has
in some cases been corroborated by historical testimony, it has been
possible to demonstrate that dense chaparral once covered extensive
areas which are now grassland or under cultivation. For example,
the floor of the Santa Clara Valley southeast of Palo Alto, which is
to-day one of the great orchard regions of the State, was less than
half a century ago solid chaparral. The bare, grassy hills nearby,
with their thin young growth of oaks, were similar. Scattered but
full-sized individuals of Adenostoma along fences and occasional
patches of uncleared ground suggest this, but in the present case
we have direct historical evidence of the fact.

Mr. G. F. Beardsley, a retired mining engineer now living in
Carmel, California, spent several years of his boyhood in this vicinity,

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 79

and is able to recall with considerable accuracy the general character
of the vegetation at that time. About 1870, according to Mr.
Beardsley, the present orchard land between Palo Alto and San Jose
was solid chaparral, and patches of the scrub occurred on the floor
of the valley between San Jose and Gilroy. A demand arose in
San Francisco for the massive woody "roots" of the shrubs for fuel,
which made it profitable to clear the chaparral from valley and foot-
hills and ship it to the city. Later the value of the land for fruit and
grape culture was discovered, which resulted in the almost complete
destruction of the brush. Those cleared areas which are not under
cultivation are to-day typical California grassland, sometimes with
young oaks scattered through them.

A very interesting and significant bit of relict evidence is the
occurrence of a typical area of chaparral well out in the Sacramento
Valley. This locality, in Colusa County, between Hershey and
Arbuckle, was brought to my attention by Dr. H. M. Hall, of the
University of California. Here are several sharply defined patches,
typical remnants. One occupies a shallow draw running through
a cultivated field; another, in an uncultivated piece of level ground,
ends abruptly at a fence. In all, Adenostoma is by far the most
abundant species; it is of very large size (3 meters) and apparently
perfectly content with its home. Other species seen are Arctosta-
phylos manzanita, A. tomentosa, Heteromeles arbutifolia, and Rhamnus
crocea. Some scattered trees of Quercus douglasii grow where the
chaparral has been partially destroyed. These localities are 5 to
8 km. distant from the foothills of the Coast Range.

A few words should be added concerning the trees which occur in
scattering growth with the grasses. These are deciduous oaks,
especially Quercus douglasii, and Pinus sabiniana. Typical stands
of young Q. douglasii have been seen where it is certain that chaparral
was formerly in control. In some places remnants of chaparral
are mingled with the oaks. The habits of this species, moreover,
fit it admirably for such a life. It is an abundant seeder, withstands
severe drought, and is one of the least shade-tolerant of the Cali-
fornia trees. Pinus sabiniana is similar in habitat requirements
and frequently occurs in chaparral, especially with Adenostoma.
It is at least a reasonable supposition that the thin woodlands of
Quercus douglasii and Pinus sabiniana which cover great areas in
the Coast Ranges and Sierra foothills, often forming a wide zone
between the grassland of the plains and the chaparral of the moun-
tains, are secondary, occupying areas which would support climax
chaparral but for agencies of disturbance.

It remains to account for the absence of the chaparral in these
areas of grassland, where it is hypothetically the climax. Several
agencies have operated to bring this about, some of which have

80 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

already been suggested. Clearing for firewood has been locally
effective; clearing for cultivation much more so. By far the most
important cause of destruction has been fire. It has been stated
that fire favors the extension of the chaparral at the expense of the
forest. It is also true that fire, if it occurs with great frequency,
favors grassland at the expense of the chaparral. A single burning
of chaparral will result merely in a crop of stump sprouts and greater
density than before, but yearly burning will inevitably destroy the
brush completely or prevent invasion by it. Cattlemen and sheep-
men in the early days, according to unpublished Forest Service
reports, were accustomed to fire the brush annually in the foothills to
destroy it and thereby improve the grazing conditions. This resulted
in a great increase of grassland at the expense of the chaparral.
Such recent events, however, are of small importance compared
with the effects produced by the aboriginal population. The
following quotation from Jepson (47) is of interest in this connection :

"The herbaceous vegetation [in the Great Valley] in aboriginal days grew with
the utmost rankness, so rank as to excite the wonderment of the first whites. ■ • • •
This dense growth was usually burned each year by the native tribes, making a quick
hot fire sufficiently destructive to kill seedlings, although doing little injury to estab-
lished or even quite young trees."

Dr. Jepson writes concerning the sources of his information :

"The statement made in the Silva re periodical burning rests upon evidence gath-
ered by myself from members of the Nyah, Hupa, Porno and other tribes; also from
verbal relations of early Californians."

It may be added that Merriam estimates the Indian population
of California west of the Great Basin at its maximum to have been
250,000. Here we have suggested the cause of destruction of the
chaparral, or the prevention of its establishment. It is easy to
conceive of the possibility of an occasional area escaping disaster,
which would account for the fragments near Hershey and others like
it. The reason for the scattering woodland of Quercus douglasii
and other species is also apparent; an occasional year without burning
or a succession of such years would permit the successful germination
of young trees, which once established would resist the attacks of
fire. The patchy transition between grassland and chaparral is
also explained, for fires started in the valleys, where most of the
Indian population lived, would spread into the surrounding ranges,
in various directions and to varying distances. Certain areas would
escape, and these would be larger and more numerous toward the
interiors of the mountain systems, where paucity of population
would reduce the starting of fires to a minimum. The reasons for
the burning I have not been able to discover.

It is certain, therefore, that extensive areas which are now domi-
nated by grasses or thin forest of xerophytic trees were formerly

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 81

controlled by chaparral or would come to support chaparral if
destructive agencies were eliminated; and the principal causes of
the destruction of the chaparral or prevention of its establishment
have been clearing and repeated fires.

As to the exact extent of former chaparral dominance we can not
set limits with any degree of assurance. Some areas are made certain
by historical testimony and others almost equally so by relict evi-
dence. We are thus led to admit the extreme probability that the
great mass of the Coast Ranges, the foothills of the Sierras, and such
minor valleys as the Santa Clara were originally dominated by broad-
sclerophylls; that areas now inhabited by grasses or xerophytic
trees, occurring as isolated patches surrounded by chaparral or
forming a more or less continuous zone upon the foothills bordering
the major valleys, are such by reason of clearing or repeated fires.
It may be necessary to except from this category the ridges of Coast
Range and Sierras that immediately surround the southern end of the
San Joaquin Valley, which, with the inclosed plains, are desertlike.

There is a possibility, further, that the northern end of the Great
Valley itself may once have been dominated by chaparral. The
evidence for this is not absolutely conclusive, but is of sufficient
weight to justify presentation. The remnant near Hershey, in
the Sacramento Valley, shows that the chaparral species are able to
grow to their maximum size and to form a solid cover under typical
valley conditions. The practical absence of typical chaparral along
the northeastern border of the valley (p. 77)— the only break of
importance in the whole circle of surrounding foothills — suggests
the probability of a former connection across the valley floor. The
climax chaparral has apparently been pinched out between the
invading mass of the fire-favored grassland and the relatively resistant
barrier of the conifer forest.

So far as available data show, there are no constant efficient
climatic differences between the areas dominated to-day by chaparral
and the Sacramento Valley. Referring to table 1 (p. 19) we find
that 44 stations in the southern Coast Ranges, southern Sierra
foothills, and Cuyamaca Mountains, which are the regions most
certainly controlled by chaparral, give an average of 21.67 inches of
precipitation, while 23 stations in the Sacramento Valley average
21.85 inches. Neither is the seasonal distribution notably different,
the proportion of total precipitation occurring in the months May
to October being, in the case of the chaparral regions, 11.6 per cent,
and in the Sacramento Valley 13.9 per cent. Available temperature
data are unsatisfactory, as there are no records of mean summer
maxima for the Sacramento Valley.

Clements (22, p. 150) has advanced the theory that the Great
Valley of California, "from Bakersfield to Mount Shasta and from
the foothills of the Sierra Nevada and Cascade Mountains, through

82 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

and over much of the Coast Range" was originally grassland of the
bunch-grass type, Stipa being the most important genus, which has
been almost entirely replaced by the introduced annual grasses
(Avena fatua and A. barbata) that are so prominent to-day. He
admits the unlikeness of the climate, especially in seasonal distribu-
tion of precipitation, to that of other regions of grassland climax,
but rightly affirms that the climax plant community is itself the
best available indicator of climate. The decision rests, therefore,
upon correct identification of the climax community. Clements's
evidence of grassland dominance is based upon assumed relicts,
and my evidence of chaparral dominance is of the same nature.
In an area where individuals and patches of both types occur today,
the weight of probability would, in my opinion, favor the chaparral
plants as being the survivors of the true regional climax. It has
already been stated that most of the species of the climax chaparral,
notably Adenostoma, reestablish themselves with the utmost difficulty
once their mass control has been destroyed, and that therefore thor-
oughly isolated individuals or patches of such plants are almost
certainly relicts of former dominance rather than centers of recent
colonization. The same is not true in the case of grasses, even the
perennial species which make the grassland climaxes. With these,
greater mobility and ease of germination bring about a wide scattering
of individuals, and the acceptance of isolated plants or patches as
true relicts is therefore rather hazardous.

The acceptance of my conclusions as to former control of certain
regions (the main mass of the Coast Ranges with its minor valleys,
the foothills of the southern Sierras, and perhaps the northern end
of the Sacramento Valley) by chaparral does not necessarily preclude
the possibility of climatic grassland over much of the Great Valley.

In southern California the contact between chaparral and coastal
sagebrush resembles that between the former and the grassland in
the central part of the state. The sagebrush, which is undoubtedly
climax in certain portions of the interior valleys, has extended its
area of dominance at the expense of the chaparral, because of fire.

In conclusion, then, the climax chaparral has transgressed its
normal climatic limits along its mesophytic border through its
invasion of the forest, fire being the causative agent; on its xerophytic
border it has been pushed back a considerable distance by the grasses
and xerophytic forest in the north, and by the coastal sagebrush
in the south.

SUCCESSIONS LEADING TO THE CHAPARRAL CLIMAX.

Primary Successions.

The areas where primary succession may be observed are limited,
the bulk of the area having reached its final stage or being in process
of secondary development. Naturally most of the successions are

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 83

of the xerarch class. Three distinct lines of development have been
observed upon primary areas, corresponding with three common
modes of soil formation. These are the successions on rock surfaces,
on washes and alluvial fans, and on coastal dunes. In dealing with
such an extensive and varied region the treatment can not be com-
plete, and there is here a fruitful field for further study.

Development on rock surfaces. — Though most of the chaparral-
covered mountains have a uniform, unbroken mantle of vegetation,
there are locally extensive areas where bare rock is being invaded.
Such successions are much the same in character wherever they occur.
The usual lichens and xerophytic mosses cover the surfaces and
their establishment is assisted by the abundant scaling of the rock
due to sudden temperature changes. Crevice plants as usual are
of supreme importance. Among the herbaceous species, Pellcea
mucronata and Selaginella bigelovii are frequent. The most impor-
tant pioneers are the chaparral shrubs themselves, the successional
stages thus being greatly telescoped. Adenostoma is perhaps most
important among these. Yucca whipplei is common in such places
in the southern part of the State. Dead remains of shrubs anchored
in the crevices show that permanent establishment may be a matter
of many generations. Gradually the rock disintegrates and the
chaparral shrubs increase in number. It is not uncommon, however,
to find a solid cover of chaparral concealing and growing entirely
from the crevices of a rock-layer whose surface is barely at all dis-
integrated. Thus, through a process long in point of time but com-
prising few stages, the climax community comes to control the bare
rock areas.

Development on alluvial fans and washes. — Because of frequent
production of fresh surfaces due to erosion and deposition by periodic
torrents, there is excellent opportunity for the study of such succes-
sions. Alluvial fans and continuous piedmont slopes occur through-
out the State, but they are especially well developed in the Los
Angeles and San Bernardino Valleys, bordering the high ranges to
the north. The physiographic processes involved in their formation
are a complex combination of erosion and deposition. The periodic
torrents from the mountain canyons and ravines are continually
changing their courses, depositing material now here, now there,
cutting into deposits already made, and forming new surfaces through
both processes. If the cutting goes deep enough terraces are formed
whose surfaces are thereafter free from further direct disturbance.

Certain ancient fans now in process of dissection show all stages
in the physiographic cycle. There is a very remarkable example
south of San Timoteo Canyon which may be viewed to excellent
advantage from Smiley Heights, near Redlands. V -shaped ravines,
cut sharply into the original smoothly sloping surface of the fan,

84 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

represent extreme youth. Wider ravines with graded floors approach
maturity, and extensive base-leveled areas with occasional monad-
nocks have reached old age. With such a series it is an easy matter
to correlate vegetational development with physiographic age.
Invariably upon the flat surfaces of the terraces and uneroded
remnants the vegetation is well-developed chaparral, in which
Adenostoma is commonly dominant, either solid or represented
by relict masses. Chaparral covers the newer base-leveled areas
also, except where streams have recently coursed over them. In
other words, the stable areas, which have had a chance for a relatively
long period of vegetational development, support the chaparral
community. Further, the composition of the vegetation of oldest
terrace-tops and youngest base-levels is practically identical. Since
the terrace-tops, with their vastly greater age, have not advanced
beyond the stage reached by the recent base-levels, it is plain that
that stage is the climax.

The species that inhabit the unstable eroding slopes and the
recently formed washes make a very different assemblage. They
are mainly short-lived half-shrubs which germinate readily and are
able to exist in the unfavorable conditions which such habitats offer.
Clements (22) regards this well-marked community as the western-
most association of the sagebrush formation, because of the impor-
tance in it of Artemisia californica, and gives it climactic rank.
His appellation is here tentatively accepted. The area of coastal
sagebrush dominance has not been accurately determined, but
roughly it covers the plains and low interior valleys of southern
California from the Santa Ana Range to the San Jacinto Mountains.
Within the climax chaparral it is successional, occurring in both the
primary and secondary series. A number of the species extend
northward for varying distances, a few as far as the San Francisco
Bay region.

A list of the most important species follows, those particularly
characteristic of the southern California group being marked with
an asterisk (*) :

Eriogonum fasciculatum.* Ramona stachyoides. Baccharis pilularis.

Syrmatium glabrum. Sphacele calycina. Encelia farinosa.*

Helianthemum scoparium. Diplacus glutinosus. Eriophyllum confertiflorum.

Malacothamnus fasciculatus.* longiflorus. Lepidospartum squamatum.*

Ramona clevelandi.* puniceus.* Artemisia californica.

nivea.* Ericameria pinifolia.* Senecio douglasii.

polystachya.*

Two or three brief descriptions of individual localities will be of
interest. The eroding southwest slope of Smiley Heights, apparently
a remnant of an ancient fan, near Redlands, is steep and unstable.
The species are Eriogonum, Encelia, Ramona stachyoides, and Arte-
misia. The bluff is much dissected into gullies and ridges which

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 85

present alternating southeast and northwest slopes, and there is
sharp vegetational alternation corresponding with the slope-exposure
differences. The first two species occupy the slightly moister north-
west slopes in relatively close formation, and the last two make a
sparse covering on the opposite hot, dry exposures. The alternation
is very conspicuous when viewed from a distance. There are thus
degrees of xerophytism among the coastal sagebrush species, giving
rise to additional successional stages where delicate differentiation
of habitats exists. Remnants indicate that the original vegetation
of the level summit of Smiley Heights was Adenostoma chaparral.

Three successional stages were revealed by study of the dry
washes of Mill and Santa Ana Creeks, at the southern base of the
San Bernardino Mountains, and of other similar localities. The
plants upon the areas which had most recently been water-swept
were Baccharis viminea, Salix argophylla, S. lasiolepis, and Populus
fremoniii, suggesting the inception of a stream-bank community,
which can survive only as long as the stream holds its course. The
second group, a few individuals of which enter with the first, are
the coastal sagebrush species. The pioneer of this stage is the almost
leafless composite Lepidospartum squamatum. Others gradually
arrive, until a typical half-shrub community is established. Most
important on these washes are the following:

Syrmatium glabrum. Ramona stachyoides. Opuntia covillei.
Eriogonum fasciculatum. polystachya. bernardina.

Artemisia californica. Malacothamnus fasciculatus Yucca mohavensis.

Ericameria pinifolia. splendidus. Encelia farinosa.

Senecio douglasii.

The last is confined to the region from San Bernardino to Elsinore,
so far as southern California west of the deserts is concerned, and
Opuntia covillei is replaced by 0. occidentalis farther west; otherwise
the list is representative of the whole. Selaginella bigelovii covers
large areas of cobbles and sand with dense growth. The third and
final stage comes with the establishment of Adenostoma. In the
younger areas we find it scattered, with many of the above list mingled
with it. On the areas longer undisturbed, as well as on the higher
terraces nearby, it forms an almost pure growth. Frequently an
uneroded island is seen, surrounded by bare water-swept cobbles.
Such an area is likely to possess a remnant of pure Adenostoma.

In central California, washes, fans, and terraces are not such
striking topographic features, perhaps because the mountain fronts
are less abrupt and torrential erosion is therefore less powerful.
Still, they do occur, and the same developmental stages and the same
Adenostoma climax are present, though the successional species
are somewhat different.

Development on coastal dunes. — A comprehensive study of the
strand vegetation of the Pacific Coast is now in progress. A very

86 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

brief sketch must here suffice. The pioneers upon beach and dune are
mainly strand-succulents, such as Abronia latifolia, A. maritima,
A. urnbellata, Mesembryanthemum cequilaterale, Convolvulus soldanella,
and Franseria bipinnatifida. These are followed by a group of
half-shrubs, including Eriogonum parvifolium, Lupinus chamissonis,
Ericameria ericoides, and Eriophyllum stcechadifolium. In the
Monterey region, on an area of very ancient dunes, the succession
may be traced farther. Adenostoma and Arctostaphylos pumila follow
the half-shrubs and are themselves succeeded by Arctostaphylos vestita
and other relatively less xerophytic chaparral species. This commu-
nity is displaced by a low forest of Quercus arjrifolia, to which, on
the Monterey Peninsula, is added Pinus radiaia. This is the local
climax, more mesophytic in type than in the regions adjacent,
probably because of more favorable moisture conditions.

Secondary Successions.

Two classes of secondary successions may be distinguished — those
after occasional burning or clearing (where the basal portions of the
shrubs are left to sprout) and those after thorough destruction of
the original vegetation, either by the grubbing out of the underground
parts or by burning at very frequent intervals.

In the first case the succession is a short one, bringing a return
practically to the original state in a very few years. The composition
of the new chaparral is for a time somewhat different from what it
was formerly because of the persistence of some of the plants which
came in after the fire. The following outline gives the general features
of the development as determined by a quadrat study near Palo
Alto and observation elsewhere. Most of the species sprout readily
from the stump, and the new shoots grow with astonishing rapidity,
even in the driest part of the dry season. Six weeks after the Ojai
Valley fire of June 1917, sprouts were found which had attained a
height of nearly a meter. The sprouts therefore have a tremendous
advantage over any seedlings that may start, especially if the fire
occurs early in the dry season, since germination can not take place
until the rains begin. With the first rains, seeds which have blown in
or lain dormant germinate in enormous number and great variety.
Among these are many of the shrub species, which gradually decline
in number as the years pass. In a permanent quadrat 5 meters
square, near Palo Alto, 562 shrub seedlings appeared during the
first rainy season. Seven years later the number of survivors (in-
cluding some new germinations) was 189. Of the first arrivals, 13
per cent were Adenostoma and 62 per cent Ceanothus cuneatus,
although the original growth was nearly pure Adenostoma. The
greatest number of herbs appeared during the second rainy season.
The 2,841 recognizable individuals comprised 28 species, of which all
but half a dozen were annuals. Of this large number, only 13

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 87

individuals survived until the following November. There is
thus a yearly crop of annuals, smaller each season because of com-
petition with plants of greater stature. The shrub seedlings suffer
also in competition with the sprouts, which finally dominate the
area. A few of the seedlings survive, however, especially those of
Ceanothus, and form an important part of the stand for a number
of years. Ceanothi of several species, when present in large numbers,
seem to be indicators of recent disturbance — in other words, chaparral
with a large proportion of Ceanothus is likely to be a penultimate
stage in a secondary succession.

When the chaparral is utterly destroyed, whatever the cause,
the course of development leading to its reestablishment is usually
an exceedingly long one. Often a community of a very different
stamp takes possession, which, if the disturbing agency frequently
repeats its work, may last indefinitely and simulate a true climax.
I have already stated my conviction that the grassland of central
California is in part a stage in a secondary succession within the region
of the chaparral climax, and that certain trees which habitually
grow with the grasses, notably Q. douglasii, have a role in the same
process. If such grassland be kept moderately free from grazing
and fire, a rough, thick growth of shrubs and half-shrubs often occu-
pies it. These are Rhus diversiloba, Sphacele calycina, Diplacus
glutinosus, Baccharis pilularis; and they may represent a further
stage in the reestablishment of the climax.

In southern California the coastal sagebrush plays an even more
important part in the secondary successions than in the primary.
Great areas of it cover the mesas and lower foothills, and its spatial
relation to the chaparral is usually exactly analogous to that of the
grassland of central California. There is the same patchy alternation
between the two, with decrease of the half-shrubs and increase of
chaparral as one penetrates a mountain mass. No relation to at-
mospheric, soil, or slope factors can be made out, and the inevitable
conclusion is that it has replaced the chaparral because of the destruc-
tion of the latter by repeated fires which, starting in the valleys, the
centers of human habitation, have spread varying distances into the
foothills. Here again we find the chaparral following the forest into
the mountains, itself followed by the coastal sagebrush, fire favoring
the more xerophytic community in each case. The most important
species in the secondary coastal sagebrush are Eriogonum fascicu-
latum, Ramona nivea, R. stachyoides, and Artemisia calif ornica.
Often they are mixed, but sometimes one encounters great stretches
of a single species almost pure. The last two are characteristic
secondary species as far north as Monterey Bay. In the valley of
the Carmel River I found a deserted vineyard that was thickly
grown up to these two species — excellent evidence of their role in
secondary development.

88 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

VI. ECOLOGICAL CHARACTER OF THE BROAD-SCLERO-
PHYLL SHRUBS AND TREES.

As a text for this chapter I will quote a paragraph from the pioneer
of physiological plant geography. He is characterizing the broad-
sclerophyll type in general (80, p. 510).

"The trees are usually low, their stems generally massive, and the branches
gnarled. The leaves are at most of moderate size, about as large as the leaves of
laurel or oleander, usually smaller, or even very small; they are scarcely ever com-
pound, as a rule narrow, lanceolate or linear to acicular; their margins are usually
entire. The leaves are not generally placed with their flat surfaces perpendicular to
the strongest light, but usually avoid it by assuming an oblique or parallel position.
They are either destitute of an air-containing tomentum, or this is confined to their
under-surface; on the other hand glandular hairs are not uncommon on both leaf-
surfaces. Even when there is no tomentum the leaves comparatively speaking are
seldom shiny, but more frequently, even if smooth on the surface, are dull, perhaps
owing to exudations of resin, and often bluish. Histologically the foliage is character-
ized by the thickness of the walls of all the cells, including even the parenchymatous
ones, by the abundance of sclerenchyma, by the strong development of cuticle, and
by the diminution of the intercellular spaces; these qualities in the aggregate give the
leaf its characteristic, stiff, leathery consistency."

In the large, this is an adequate brief characterization of the
California broad-sclerophylls. The following detailed analysis will
confirm Schimper's statements in the main, at the same time pointing
out features in which the California species diverge from his generali-
zations.

GROWTH FORM.

A very great majority of the California broad-sclerophylls (88
per cent) are shrubs. In Raunkiar's scheme (83) they would be
classed as nanophanerophytes. The height of mature individuals
is naturally variable. Adenostoma may be found, under various
conditions, 3 dm. high to 4 meters high. The average for all shrubby
species, for the region as a whole, may fairly be placed between
1.5 and 2 meters. In a single locality the height of the individuals
of all species is usually remarkably uniform. A few species, which
do not usually occur in abundance, habitually rise above their
neighbors. Such are Heteromeles arbutifolia and Garry a elliptica
and those which are potential trees. The broad-sclerophyll trees
come mainly under Raunkiar's class of mesophanerophytes (8 to 30
meters). Castanopsis occasionally attains a greater height, and
probably Pasania and Arbutus also. Most of the trees reach great
size under favorable conditions. The diameters of the largest speci-
mens that I have measured are as follows: Quercus agrifolia, 2.1
meters; Quercus chrysolepis, 1.8 meters; Arbutus menziesii, 1.5 meters.
Greater diameters have been reported for these species, and Sargent
(79) reports 3 meters as the maximum for Castanopsis, 1.8 meters
for Pasania, and 1.5 meters for Unibellularia.

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 89

The branching of the shrub species is notoriously close and intri-
cate, and the clump habit is almost universal, partly but not wholly
due to fire. Many of the tree species show a similar manner of
growth— several stems from a single base (often because of fire) — and
contorted assurgent or horizontal branches. Quercus agrifolia is a
notable instance, with its branches of enormous size frequently
resting on the ground. Most of the trees are marked by great
spread of branches rather than by height. A specimen of Quercus
agrifolia near Palo Alto shades an area 30 meters in diameter, and
this is not an extreme case. The deliquescent mode of branching
is the rule. Even when there is a distinct trunk, it is soon lost in
the branches. Pasania and Castanopsis exhibit the closest approach
to the excurrent type, maintaining a distinguishable trunk almost
throughout. Young specimens of Umbellularia show the same form
very perfectly, so that one frequently takes them at a distance to
be conifers. Large trees of this species are usually branched near
the base, but the separate stems maintain the excurrent form.

This seems an appropriate place to mention the small but con-
spicuous group of species which have more or less spinescent branches.
These are Xylothermia montana, Ceanothus cordulatus, C. divaricatus,

and C. incanus.

ROOT SYSTEM.

Somewhat scanty evidence indicates that the root systems of
the chaparral shrubs, when growing in loose soil, are prevailingly
of the "dual type" described by Cannon (19). Such root systems
are in part deeply penetrating and in part superficial, the latter
making up the greater bulk. Cannon states that the sclerophyll
Quercus agrifolia and the deciduous Q. douglasii and ALsculus cali-
fornica are of this type, and suggests that the deeply penetrating
roots, which may reach to the neighborhood of the water-table,
are formed in youth, while the more numerous superficial roots
are developed later.

The root system of a large specimen of Adenostoma growing on
Jasper Ridge near Palo Alto was carefully excavated and charted
(fig. 19). This plant grew in a spot where the soil was pure sand,
gradually compacted downward into undecomposed sandstone.
It was impossible to trace the roots to a greater depth than a meter,
and this was the only locality on the ridge where excavation was
feasible at all. The plant grew close to station 10 of the habitat
study series. Its root system was decidedly of the dual type.
Figure 19 (lower), drawn as if the roots grew all in one plane, exhibits
their vertical distribution. It is very common for a large root, at
first horizontal, to bend suddenly downward. Several such, lost at a
depth of a meter, were 6 to 7 mm. thick at that point. No tap-root
or anything approaching it was found, though every seedling possesses

90 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

a well-developed one. Small branches of 1 mm. or less diameter
were given off rather regularly, but at long intervals, all along the
main roots. These terminated in groups of mycorhizal rootlets
which were embedded in masses of sand-grains held loosely together
by hyphae.

Fig. 19. — Root system of Adenostoma fasciculatum: upper drawing shows hori-
zontal, lower shows vertical distribution. In the latter case the roots
are arbitrarily drawn as if growing in one plane.

Another excavation of considerable extent was made nearby to
learn something of the subterranean relations between the chaparral
shrubs and the lower plants that accompany them. The dominants
here were Adenostoma fasciculatum, Quercus durata, and Arctostaphylos
tomentosa. The individuals grew rather far apart, and therefore
the accompanying species were more numerous than usual. A second
stratum included Eriodictyon californicum, Rosa californica, Helian-
themum scopariu?n, Diplacus glutinosus, and Syrmatium glabrum,
low shrubs and half-shrubs. A third stratum comprised Micromeria
chamissonis, Gymnogramme triangularis, and a few grasses. Below
the surface only two strata were distinguishable. The upper one,
about 20 cm. in depth, with some humus, was fairly well filled with
the roots from the second and third aerial strata, together with great
numbers belonging to the dominants, which were most conspicuous
because of their large size. Below 20 cm. were found only the scat-
tered, deeply penetrating roots of the dominants.

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 91

General observation shows that the three principal species of
Jasper Ridge, Adenostoma, Quercus, and Arctostaphylos, all possess
the dual root system. Its usefulness is apparent. During the winter
and spring, when most of the year's growth is accomplished, the
abundant superficial roots absorb large quantities of water. The
humus in this stratum both increases water-retaining capacity and
adds to the supply of other food materials. During the dry summer
the deeply penetrating roots supply the plant with the minimum
amount of water necessary to tide it over the critical period. It is
inevitable, when one recalls the extreme desiccation of the surface
soils, that the superficial roots cease functioning during the dry
season. The dual system is of course apparent only in plants growing
in ordinary soil. Where the rock is at the surface, which is very fre-
quently the case, the roots penetrate the crevices wherever they can,
without observable system.

Eriodictyon californicum has a very different root plan (fig. 20).
The main roots are few, and they travel close to the surface, often
from 1 to 6 cm. below it, and occasionally one turns abruptly down-

Fig. 20. — Root system of Eriodictyon californicum. For explanation, see fig. 19.

ward to the deeper soil-layers. From the horizontal portions
sucker shoots appear, so that a number of apparently independent
plants are often found to be connected below ground. Whenever
the roots are exposed, as by erosion, they produce shoots abundantly.
The roots of Eriodictyon are useful, therefore, for propagation as well
as for absorption.

The three principal chaparral species of Jasper Ridge, Adenostorna
fasciculatum, Quercus durata, and Arctostaphylos tomentosa, have
been proved to be mycorhizal. In the excavations no normal root-
hairs were seen, but it is hardly reasonable that the most deeply
penetrating roots should be associated with fungi. The mycorhizas

92 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

are most abundant naturally in the upper soil layers, where there is
most humus. Pockets of nearly pure humus were occasionally
found, probably derived from rotted roots, which were thoroughly
penetrated by the coralloid rootlets. The mycorhizas were by no
means confined to the regions of abundant humus, however. The
roots which penetrated the pure sand below the humus layer were
almost as well supplied with the fungi. In station 2, where the soil
is merely a slightly decomposed clayey sandstone, well-developed
mycorhizal roots of Arctostaphylos were taken from tight crevices
between the fragments.

The best material for study was obtained from Quercus durata.
The last few centimeters of the roots were found to branch freely,
giving off numerous clumps of blunt-tipped coralloid rootlets, each
clump firmly embedded in a mass of sand-grains bound together by a
network of fungal hyphse. The hyphse were of two kinds: (1) very
small, transparent, branched, quite certainly non-septate, much
more abundant than (2), doing most of the sand-binding; (2) diameter
twice or thrice that of (1), branched, very dark brown, some opaque,
others only partially so, plainly septate. What appeared to be
tight bundles of hyphse of the septate type were seen, simulating
rootlets. A cross-section of a mycorhizal root of this species is
given in plate 20a. The dense felt of mycelium completely sheathing
the root is well shown, and the hyphse are seen to penetrate between
the cells of the cortex, so that some of the outer ones are apparently
isolated from their neighbors. The strands which originally extended
out into the soil have, of course, been lost in the process of slide-
making.

THE LEAF.
THE DECIDUOUS ELEMENT.

Of the 91 species listed on pages 113 to 120 as being included in the
broad-sclerophyll communities as dominants, 18 or 19.8 per cent
are deciduous. These are as follows (two or three species which lose
their leaves in late winter or spring are classed as evergreens) :

Corylus rostrata californica.* Amelanchier alnifolia.* Acer macrophyllum.*

Quercus brewed. Prunus demissa.* ^Esculus californica.

douglasii. emarginata.* Ceanothus integerrimus.

garryana.* subcordata. parryi.

kelloggii. Cercis occidentalis. parvifolius.

lobata. Acer glabrum.* sanguineus.*

Omitting those which are not characteristically Californian
(marked by asterisk), the number is reduced to 10, or 11 per cent.
The relation of the evergreen habit to the Californian habitat is
further shown bjr a comparison of the three communities. The
climax chaparral, most characteristically Californian, with a total
of 44 dominant species, has no deciduous dominants, the broad-
sclerophjdl forest, with the small total of 13, has 4 deciduous, or

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 93

30.8 per cent, the conifer forest chaparral, with 29 dominants, has
12 deciduous, or 41.4 per cent. Of these 12, 7 are not characteristic-
ally Californian, ranging far beyond the boundaries of the State.
It is thus evident that the most characteristically Californian com-
munity is made up entirely of evergreens, so far as its dominants are
concerned, and that the community which has the greatest proportion
of deciduous species is also the least characteristically Californian.
Further discussion of the leaf will be confined to the evergreen species,
or, in other words, to the broad-sclerophylls.

EXTERNAL CHARACTERS.

Compound leaves. — Of the 74 broad-sclerophyll species, only 2
have compound leaves. These are Berberis pinnata and Xylothermia
montana. The leaves of the first are pinnate, with 5 to 9 leaflets,
and the second palmate, ordinarily with 3 leaflets. In the measure-
ments to follow, the individual leaflet of these species has been taken
as the unit.

Form. — Lobing is still rarer. Fremontodendron californicum is
the only species, and even in this many of the leaves are unlobed.
In shape there is great uniformity. Of the 71 species which are
neither compound nor lobed, 54 are of the oval tj^pe, ranging from
orbicular through elliptic to ovate, the great majority being elliptic.
Ten are distinctly obovate, 5 are lanceolate or oblanceolate, and 2
(genus Adenostoma) are linear and terete. Adenostoma fasciculatuw. ,
the most important single species, is not a frroad-sclerophyll at all,
but, as already suggested, this is hardly sufficient ground for dis-
carding for the whole type that very satisfactory appellation.

Size. — It is well known that leaf-size is a fairly trustworthy measure
of habitat, especially of the moisture element. Raunkiar has pro-
posed an ecological classification of plants upon the basis of this
character, which has been recently brought to our attention in the
translation by Fuller and Bakke (31). He has established several
size-classes based upon the surface area, the upper limits of which are
as follows:

sq. mm.

1 . Leptophyll 25

2. Nanophyll (9 by 25 sq. mm.) 225

3. Microphyll (9 2 by 25 sq. mm.) 2,025

4. Mesophyll (9 3 by 25 sq. mm.) 18,225

5. Macrophyll (9 4 by 25 sq. mm.) 164,025

6. Megaphyll

Some other delimitation of the classes would have served much
better for the broad-sclerophylls of California, since most of the
leaves seem to group themselves about the division-lines, but for
the sake of comparison with other regions it is better to hold to the
system that Raunkiar has established. All the leaves to be classified
here come under the first four classes, and it may be helpful to state

94 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

that a leptophyll, if of the common elliptic form, will be about 9 mm.
or less in length, and that the upper length limits of the next three
classes, assuming the same leaf -form, will be approximately 27 mm.,
81 mm., and 243 mm. The sizes of the leaves were determined by a
study of the specimens in the herbarium of the University of Minne-
sota, which has an excellent representation of the Californian flora.
To provide for cases in which a species overlaps two classes, three

36

34

32

30

28

26

io24

Id

U 22

Id

$20

fe '8

* \

or 16

/ \ \

hJ

/

\ \

* 14

/ i
I i

\ \
\ \

g«

/ /
/ /
/ /

\ \

10

■

/ *

/ /

/ i

\ \
\ \
\ \

8

/ /
/ /
J i

6

/ '

X \

> \

4

.>>'■•■. * \

y '■••. A \

2

■*»<£• — "

LN

N
Total

NMi

Mi

MiMe Me

Climax Chaparral

Conifer Forest Chaparral

Broad-sclerophyll Forest

Fig. 21. — Leaf size in the broad-sclerophyll communi-
ties. See text.

intermediate categories were made. Because such a large number
were found to cluster around the division-lines, it was thought best
to include also in the intermediate classes those which were con-
sistently very close to the limit, though not actually crossing it, so
far as was shown by the specimens examined.

Figure 21 gives the results of the study. The generally small size
of leaf is manifest, all but 1 1 of the 74 species falling within the three
lowest classes — leptophylls, nanophylls, and microphylls. Further-

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 95

more, there is a very decided maximum in the NMi region, practically
half of all being in this class. Species with very small leaves are
scarce, though it must not be forgotten that Adenostoma fasciculatum,
most important of all, is one of these. Considering the vegetation
by communities, we find that of 42 important species of the climax
chaparral nearly half are of the NMi class, with the rest well dis-
tributed on both sides of it. Of 17 species of the conifer forest
chaparral, those of class NMi compose nearly one-half, but the
others are all of larger classes. Of the 9 broad-sclerophyll forest
species, none are smaller than Mi and the maximum lies in the class
MiMe. There is thus shown a relation between leaf-size and habitat,
the community living in the driest habitat having the largest pro-
portion of small leaves. There is also a correlation between size of
leaf and size of plant, since the trees of the broad-sclerophyll forest
show consistently greater leaf-size than the shrubs of either of the
two chaparral communities.

Attitude. — In the majority of species the leaves lie horizontally,
but there is a small group with vertically placed leaves. Certain
species of Arctostaphylos are prominent here, especially A. glauca,
A. hookeri, and A. viscida. In many others there is a more or less
prevalent tendency toward the vertical position. Dendromecon
rigidum has leaves that are ordinarily vertical, but, as in many other
cases, they are horizontal in the more mesophytic situations.

Margin. — A majority of the 74 species (58.1 per cent) have entire-
margined leaves, although a number of these may show occasional
dentation, especially on stump sprouts. 24.3 per cent have leaves
that are more or less toothed in the normal fashion; they are serrate,
dentate, or crenate. In nearly every case the teeth are shallow. A
smaller but conspicuous and characteristic class (17.6 per cent)
have notably spiny-toothed leaves that may be aptly described as of
the holly type. The resemblance in leaf character between the
various species of this group is often striking. This is particularly
true of Prunus ilicifolia and Rhamnus crocea, whose leaves are
frequently so similar as to require very minute examination to dis-
tinguish them. Eight species (10.8 per cent) have leaves with notably
revolute edges. This character appears inconstantly in a number of
others.

Pubescence. — Pubescence is not a prominent feature of the broad-
sclerophylls. Of the 74 species, 48.7 per cent have more or less
pubescent leaves, but in only 17.6 per cent are thej r pubescent on
both surfaces, and often the covering is sparse. The leaves with
hair-filled cavities are not considered here unless the hairiness is
evident to superficial examination.

Miscellaneous. — Two species, Eriodictyon californicum and Ceano-
thus velutinus, have leaves that are varnished and glutinous on the

96 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

upper surface. In both they are pubescent below. A few species of
Ceanothus and Arctostaphylos have more or less sticky-glandular
leaves. Three species commend themselves to the sense of smell —
Myrica calif ornica, Umbellularia californica, and Eriodictyon cali-
fornicum. The last two are widespread and common, and thus give
the broad-sclerophyll vegetation a reputation for odoriferousness
which it hardly deserves.

The external characteristics of the 74 species are presented in
detail in table 12.

Table 12. — External leaf character.

Corn-pound (2.7 p. ct.):
Berberis pinnata.
Xylothermia montana.
Lobed (1.4 p. ct.):

Fremontodendron califor-

nicum.
Size:

Leptophylls (4.1 p. ct.):

Adenostoma fascicu-
latum.

A. sparsifolium.

Ceanothus dentatus.
Nano-leptophylls(2.7p.ct.) :

Ceanothus foliosus.

C. rigidus.
Nanophylls (9.5 p. ct.):

Cneoridium dumosum.

Ceanothus cuneatus.

C. pinetorum.

C. prostratus.

C. verrucosus.

Arctostaphylos myrti-
folia.

A. nummularia.
Nano-microphylls (45.9
p. ct.):

Quercus dumosa.

Q. durata.

Q. vaccinifolia.

Cercocarpus betulse-
folius.

C. ledifolius.

Xylothermia montana.

Rhamnus crocea.

Ceanothus cordulatus.

C. crassifolius.

C. divaricatus.

C. diversifolius.

C. hirsutus.

C. incanus.

C. megacarpus.

C. papillosus.

C. sorediatus.

C. spinosus.

C. thyrsiflorus.

C. tomentosus.

Arctostaphylos drupacea.

A. glauca.

A. hookeri.

A. manzanita.

A. mariposa.

A. montana.

Size— Cont'd.
Nano-microphylls (45.9

p. ct.) Cont'd.
A. nevadensis.
A. parryana.
A. patula.
A. pumila.
A. pungens.
A. stanfordiana.
A. tomentosa.
A. vestita.
A. viscida.
Microphylls (20.3 p. ct.):
Castanopsis semper-

virens.
Quercus chrysolepis.
Q. engelmanni.
Q. wislizeni.
Berberis pinnata.
Dendromecon rigidum.
Prunus ilicifolia.
Rhamnus californica.
Ceanothus palmeri.
Fremontodendron cali-

fornicum.
Garrya elliptica.
G. fremontii.
Comarostaphylis diver-

sifolia.
Xylococcus bicolor.
Arctostaphylos ander-

sonii.
Micro-mesophylls (14.9 p. ct.)
Myrica californica.
Castanopsis chryso-

phylla.
Quercus agrifolia.
Pasania densifiora.
Umbellularia californica.
Heteromeles arbutifolia.
Rhus integrifolia.
R. laurina.
R. ovata.

Ceanothus velutinus.
Eriodictyon calif ornicum.
Mesophylls (2.7 p. ct.):
Quercus sadleriana.
Arbutus menziesii.
Margin:

Entire (58.1 p. ct.):

Castanopsis chrysophylla.
C. sempervirens.

Margin — Cont'd.

Entire (58.1 p. ct.) Cont'd.
Quercus vaccinifolia.
Umbellularia californica.
Dendromecon rigidum.
Cercocarpus ledifolius.
Adenostoma fascicu-

latum.
A. sparsifolium.
Xylothermia montana.
Cneoridium dumosum.
Rhus integrifolia.
R. laurina.
R. ovata.

Rhamnus californica.
Ceanothus cordulatus.
C. cuneatus.
C. divaricatus.
C. hirsutus.
C. incanus.
C. megacarpus.
C. palmeri.
C. spinosus.
Garrya elliptica.
G. fremontii.
Arbutus menziesii.
Xylococcus bicolor.
Arctostaphylos drupa-
cea.
A. glauca.
A. hookeri.
A. manzanita.
A. mariposa.
A. montana.
A. myrtifolia.
A. nevadensis.
A. nummularia.
A. parryana.
A. patula.
A. pumila.
A. pungens.
A. stanfordiana.
A. tomentosa.
A. vestita.
A. viscida.
Toothed (24.3 p. ct.):
Myrica californica.
Quercus engelmanni.
Q. sadleriana.
Pasania densifiora.
Cercocarpus betula?-
folius.

Species- density of broad-sclerophylls, shown by depth of shading, in California and in western North America.

,3^

3 Ml

* 70* x\ (

~\ .'s 1 xes^ — \ ~~i

/ L
X / ■=

^ / I

»■ O'-YMP/*

UlEliNA

s / f~

/ SALEM~4"

S ,20jA-.

- 11

1

G ^""Cx

r A\

m J / If

|C*R55n citV

■7 IN VA Y

IP M-J

"^ \CI

D A j

1 ^

^SftLT^AKt CITY .

Y%0 u

1 |

JtNVtR .f \

l/~* T§

1

S=!.?S\
■V s

; 1

-TOTAL PRECIPITATION

|k

4 7. O^hM

p>escott r^-.

\ K \

^^HOENIX yl

' j / .ci

¥ J i

1VL— T

Less than 10 in.
I0to30in.

. More than 30ln.

c zo^Cr \ A.

V %. /30T-~^J \ / \

s US 1 1

xJ\t e x ^f
c o l\ \_>-^v

100

200 300 *00 500 MILES

distribution of invcipihition in western Nurth America, and average total sea^jmil precipitation in region with \v
than 20 per cent summer precipitation.

Distribution of Adenostoma fascicuiatum: dots show autnentic localities; solid black, areas where it has been accurately mapped
as dominant; crosses show localities which are doubtful or where it probably does not exist at the present time.

COOPER

A. Subalpine forest zone in the northern Sierras: conifer-forest chaparral m

foreground; Mount Lassen in the distance.

B. Secondary thicket growth of Quercus kelloggii and (J. garryana, with relicts

of Pinus ponderosa and Pseudotsuga mucronata. South Fork Mountain.
Trinity County.

C. Conifer-forest chaparral association: Arctostaphylos patida, A. nevadensis,

Castanopsis sempervirens, Ceanothus velutinus, Quercus vaccinifolia, Amel-
anchier alnifolia; remnants of climax forest. Near Mount Lassen, northern
Sierras.

A. Transition between broad-sclerophyll and redwood forests: a group of redwoods on

flood-plain in left center; broad-sclerophvll forest (Pasania-Quercus-Arbutus associa-
tion) at right, with admixture of Pseudotsuga. South Fork of Eel River, Humboldt
Count v.

B. Undergrowth of Quercus agrifolia-Arbutu* association: Symphoricarpox racemosus

most abundant. Near Palo Alto. Santa Clara County.

A

A Quercus agrifolia-lobata association on alluvial fan: winter aspect, the (J. lobat., leaf-
less. Atherton, San Mateo County.
H Quercus chrysolepis consociation fsuccessional) on talus. Yosemite Valley.

A. Alternation of chaparral (Adenostoma consociation) and broad-selerophvll forest

correlated with south and north-facing slopes. The trees in the immediate foreground
are secondary oaks in a grassy area. Santa Lucia Mountains, Monterey County.

B. Alternation of chaparral (Adenostoma consociation) and broad-sclerophyll forest

(Quercus agrifolia-Arbutus association), corresponding with south and north facing
slopes; Quercus agrifolia, Q. kelloggii, Arbutus menziesii, and Umbellvlaria californica
(conical form) prominent in the latter; chaparral somewhat disturbed. Chaparral
(Arctostaphylos dominance) at the top of the north facing slope; also a clearing.
Permanente Canon, east slope of Santa Cruz Mountains, near Black Mountain.

A. Vegetation cover of lower Cuyamaca Mountains: chaparral on all slopes;

occasional broad-sclerophyll trees. Viejas Grade, San Diego County.

B. Chaparral cover of Santa Lucia Mountains: Adenostoma most abundant;

Arctostaphylos glauca (light tone) at left; Quercus agrifolia in ravine ai
right; occasional dead stalks of Yucca whipplei; secondary grassy areas.
Near Tassajara Springs, Monterey County.

C. Mountains of Ventura County, solidly clothed with chaparral. Near Wheeler

Hot Springs.

A. Chaparral of Adenostoma fasdcvlatum and Ceanothus crassifolius (light),

a common combination in southern California. Near Wheeler Hot Springs,
Ventura County.

B. Ground cover at Station 3, Jasper Ridge; relatively abundant litter beneath

Arctostaphylos.
C Station 2, Jasper Ridge, showing atmometers in position.

A. Ja •!>■!' Ridge: vicinity of Station 3 from Station t: Arcto i and

(jut reus durata.

B. South-facing slope at Jasper Elidge upon which Stations4 and 5 are located; from vicinity

of Station 3.

W^rm

#

SigP

j»^.

%£

A

A. Quercus agrifolia-Arbutus association on north slope at Jasper Ridge; Stations 7 and S

are in the area of the picture; chaparral at ridge top.

B. Typical scenerv of the interior north Coast Ranges, where the dominating vegetation

is a mixture of broad-sclerophylls and conifers, especially Psetidotsuga mucronata:
secondary grassy areas; Mount St. Helena.

A. Remnant of chaparral (Adenostoma consociation on the floor of the Sacramento

Valley, in Colusa County, south of Arbuckle; Coast Range in the distance.

B. Adenostoma consociation on cast slope of north Coast Ranges: clearing and

secondary grassland with Finns sabiniana and Quercus douglasii growing with
the latter; Pinus ponderosa on distant ridges. Beegum, Tehama County.

C. R< mnant of original chaparral cover on the slope of an isolated mountain; sec-

ondary grassland, with abundant cattle trails; live oaks in ravines. Mount
Diablo, Contra Costa County-

A. Primary succession in the climax chaparral region: lichens and mosses

on rock surface; Adenostoma, Arctostaphylos, and Yucca as crevice
pioneers. Cuyamaca Mountains, San Diego County.

B. Later stage of succession, showing the preeminent importance of the

crevices. Same locality as the last.

C. Fully developed chaparral (Adenostoma consociation) with Yucca

whipplei. Same locality as the last.

.dj%

A

Primary succession on alluvial fan: recent wash with scattered pioneers ol
numerous species; low terrace at right with partially developed Adenostoma
climax; uneroded remnant at left with Adenostoma. Mill Creek Wash, south
base of San Bernardino Mountains.

B. Primary succession on alluvia! fan: coastal sagebrush on old wash; Syrmahum

glabrum, Eriogonum fascicvlatum, and others; Adenostoma climax on terrace.
Mill Creek Wash, south base of San Bernardino Mountains.

C. Coastal sagebrush on south-facing unstable slope: Ramona stachyoides,

Artemisia californica, Eriogonum fasciculaium, Eneelia farinosa. Smiley
Heights, near Hedlands.

•it: .. ■■ .

■>-.■■ «r ■•-Jft*

A

* . s

Young stump sprouts of Adenostoma fascictdatum.

Destruction of chaparral bv the Ojai Valley fire of June 1917; underground parts
killed in many cases; edge of burned area at top of ridge. Near Wheeler Hot Springs

A. Chaparral-covered mountains north of Arrowhead Springs. San Bernardino Mountains;

fire-restricted remnants of Pseudotsuga macrocarpa.

B. Rebel patches of chaparral (Adenostoma consociation'; secondary grassland on ridge

top; secondary coastal sagebrush on slopes too steep for grazing. ° San .lose Canon.
Monterey County.

A

^

A. Quercus durata: cross-section of mycorhizal rootlet; note the thick external felt of

mycelium, and the hyphae penetrating between the cells of the cortex. X 170.

B. AdenoatomafasciLulaliun: cross-section of normal leaf. X 75.

C. Adenostoma fascicvlatum: cross-section of leaf from stump sprout. X 70.

D. Adenostoma fasciculatum: cross-section of leaf from seedling. X 70.

a

\ ) \ \t (

B

A. Adenostoma fascicvlatum: a is the normal xerophytic type; 6 is from
a shady mesophytic situation (Station 7 at Jasper Ridge).

15. Adenostoma fasciculatum: upper row, normal leaves; middle, leaves
from a stump sprout; lower, from seedlings.

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 97

Table 12. — External leaf character — Continued.

Margin — Cont'd.
Toothed — Cont'd.

Ceanothus dentatus.

C. diversifolius.

C. foliosus.

C. papillosum.

C. sorediatus.

C. thyrsiflorus.

C. tomentosus.

C. velutinus.

C. verrucosus.

Fremontodendron cali-

fornicum.
Comarostaphylis diver-

sifolia.
Arctostaphylos ander-

sonii.
Eriodictyon calif ornicum.
Spiny-toothed (17.6 p. ct.):
Quercus agrifolia.
Q. chrysolepis.
Q. dumosa.
Q. durata.
Q. wislizeni.
Berberis pinnata.
Heteromeles arbutifolia.
Prunus ilicifolia.
Rhamnus crocea.
Ceanothus crassifolius.
C. pinetorun.
C. prostratus.
C. rigidus.
Revolute (10.8 p. ct.):
Quercus durata.
Cercocarpus ledifolius.
Rhamnus californica.
Ceanothus crassifolius.
C. dentatus.
C. papillosus.
Garrya elliptica.
Xylococcus bicolor.
Surface:

Glabrous (52.7 p. ct.) :
Myrica californica.
Quercus agrifolia.
Q. vaccinifolia.

Surface — Cont'd.

Glabrous (52.7 p. ct.) Cont'd.

Q. wislizeni.

Umbellularia califor-
nica.

Berberis pinnata.

Dendromecon rigidum.

Heteromeles arbuti-
folia.

Adenostoma fascicu-
latum.

A. sparsifolium.

Prunus ilicifolia.

Xylothermia montana.

Cneoridium dumosum.

Rhus integi ifolia.

R. laurina.

R. ovata.

Rhamnus crocea.

Ceanothus cordulatus.

C. divaricatus.

C. foliosus.

C. palmeri.

C. pinetorum.

C. prostratus.

C. spinosus.

C. thyrsiflorus.

C. verrucosus.

Arbutus menziesii.

Arctostaphylos drupa-
cea.

A. glauca.

A. hookeri.

A. manzanita.

A. myrtifolia.

A. nevadensis.

A. nummularia.

A. parry ana.

A. patula.

A. pungens.

A. stanfordiana.

A. viscida.
Lower surface pubescent
(31.1 p. ct.):

Castanopsis chryso-
phylla.

INTERNAL STRUCTURE.

Surface — Cont'd.

Lower surface pubescent

(31.1 p. ct.) Cont'd.
C. sempervirens.
Quercus chrysolepis.
Q. dumosa.
Q. engelmanni.
Q. sadleriana.
Pasania densifiora.
Cercocarpus betulae-

folius.
C. ledifolius.
Ceanothus crassifolius.
C. cuneatus.
C. incanus.
C. megacarpus.
C. rigidus.
C. sorediatus.
C. velutinus.
Garrya elliptica.
G. fremontii.
Comarostaphylis diver-

sifolia.
Xylococcus bicolor.
Arctostaphylos ander-

sonii.
A. pumila.
Eriodictyon californi-

cum.
Both surfaces pubescent

(16.2 p. ct.):
Quercus durata.
Rhamnus californica.
Ceanothus dentatus.
C. diversifolius.
C. hirsutus.
C. papillosus.
C. tomentosus.
Fremontodendron cali-

f ornicum.
Arctostaphylos mari-

posa.
A. montana.
A. tomentosa.
A. vestita.

I have examined the leaves of 26 of the representative broad-
sclerophyll species of the central Coast Ranges. 7 are trees of the
broad-sclerophyll forest and the others are shrubs of the climax
chaparral. All the important genera are well represented, and it is
reasonable to assume that the selection presented here gives a true
picture of Californian broad-sclerophyll structure. The material
was collected near Palo Alto and in the Monterey region. It was
killed and preserved in formalin-alcohol, cut in celloidin, and stained
with safranin and aniline blue. 17 species have been selected for
illustration by means of drawings executed by Miss Vinnie A. Pease,
of the University of Minnesota. In addition to the general study,

98 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

measurements were made of thickness of leaf and of upper and lower
cuticle in each species (table 13). For the sake of comparison, a
group of two sclerophyll species of the redwood-forest undergrowth
and one of 5 native deciduous trees have been added. The figures
given in the table are averages of 10 measurements, made in nearly

Table 13. — Thickness of leaf and cuticle.

Broad-sclerophyll forest:

Myrica californica

Castanopsis chrysophylla

Quercus agrifolia

Q. chrysolepis

Pasania densiflora

Umbellularia californica

Arbutus menziesii

Climax chaparral:

Quercus durata

Berberis pinnata

Dendromecon rigidum

Heteromeles arbutifolia

Adenostoma fasciculatum

Prunus ilicifolia

Xylothermia montana

Rhamnus californica

R. crocea

Ceanothus cuneatus

C. papillosus

C. rigidus

C. sorediatus

Garrya elliptica

Arctostaphylos hookeri

A. pumila

A. tomentosa

A. vestita

Eriodictyon californicum

Redwood-forest undergrowth :

Gaultheria shallon

Vaccinium ovatum

Deciduous species:

Salix lasiolepis

Alnus rhombifolia

Quercus lobata

Platanus racemosa

Acer macrophyllum

Summaiy:

1. Broad-sclerophyll forest

2. Climax chaparral

3. Average of classes 1 and 2. . .

4. Redwood-forest undergrowth.

5. Deciduous species

Leaf.

microns.
80
254
314
219
217
208
264

205
178
371
339
562
345
385
265
246
462
214
472
224
365
398
362
355
264
345

432
365

111

111
160
149
106

222
336
314
398
127

Upper
cuticle.

microns.
3.08
2.63
3.75
3.75
2.85
5.58
2.96

2.94
3.82
4.20

11.58
9.73
3.98
5.85
8.25
5.28
4.65
2.25
6.86
1.12

17.25

11.82
8.85
9.89

12.10
1.24

05

.57

35
12
06
21
16

3.52
6.36
5.69
4.31

1.58

Lower

cuticle.

microns.
1.84
1.39
1.80
2.59
2.06
3.78
1.80

3.21

2.81
2.67
7.57
9.66

2.81

68
49
17
96
08
50
12
52

11.63

5.77

11.32

11.44

1.05

2.51
1.61

6.37

7.50

12.75

12.00

6.75

2.04
5.23
4.43
2.06
.91

every case upon 10 separate leaves, except in Adenostoma fasci-
culatum, Arctostaphylos tomentosa, and Quercus durata, where the
averages are of 40 to 50 separate leaves.

The following order has been determined upon in the descriptions :
(1) general; (2) mesophyll; (3) epiderm; (4) stomata; (5) special
features.

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 99

Broad-sclerophyll Forest.

Myrica californica.-(l) The thinnest of all the leaves studied. (2) Bifacial. Pali-
sade tissue of one complete and one partial layer, occupying one-third of the mesophyll;
in some places a suggestion of palisade layer next to the lower epiderm; sponge very loose.
(4) Stomata on lower side only, slightly elevated. (5) A sheath of tannin-filled cells
surrounds the bundles; occasional peltate glands in depressions on both surfaces.

Castanopsis chrysophylla (fig. 22).— (1) The specimens examined are of var. minor, from
Monterey, and the leaves are probably somewhat more xerophytic in structure than those
of the typical tree-form of the northern Coast Ranges. (2) Bifacial. Mesophyll dense;
palisade tissue about four layers deep, occupying a little less than half the mesophyll,
composed of rather crowded oval or nearly globular cells. (3) Epiderm, no special fea-

Fig. 22. — Castanopsis chrysophylla: section of leaf . X125.

Fig. 23. — Quercus agrifolia: section of leaf. X125.

Fig. 24. — Umbellularia californica: section of leaf. X125.

25

26

Fig. 25. — Arbutus menziesii: section of leaf. X 125.
Fig. 26. — Quercus durata: section of leaf. X125.

tures; incomplete hypoderm on both sides, mainly related to veins. (4) Stomata on lower
side only, slightly elevated, surrounded by collar-like ridges. (5) Tannin in cells sur-
rounding bundles and in hypoderm; struts of mechanical tissue from epiderm to epiderm,
following veins; multicellular trichomes covering lower surface, producing a golden fuzz.
Quercus agrifolia (fig. 23).— (2) Leaf bifacial but not perfectly so. Palisade tissue about
three to four layers deep, occupying about half the mesophyll and merging gradually into
the sponge, the cells of which are mostly elongated perpendicularly to the surface, thus
possessing palisade character; one or two incomplete palisade layers on lower side also.
(4) Stomata on lower side only. (5) Struts of mechanical tissue as in Castanopsis, but
more pronounced.

100 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

Quercus chrysolepis. — The specimens are from the shrubby chaparral form. Even so,
they exhibit less of xerophytic character than the last, in that they are more perfectly
bifacial and that the sponge is more typical. Otherwise they are essentially like those of
Q. agrifolia.

Pasania de>mflora.—(2) Bifacial. Palisade tissue about two layers in depth, occupy-
ing less than half the mesophyll; sponge not abundant, most of the space being taken up
by groups of large, thin-walled parenchyma cells, apparently water-storage tissue. (3) A
complete layer of hypoderm beneath the upper epiderm and resembling it; lower epiderm
papillate. (4) Stomata on lower side only. (5) Struts of mechanical tissue like those of
the other members of the family.

Umbellularia calif ornica (fig. 24).— (2) Imperfectly bifacial. Palisade tissue two layers
in depth, occupying half the mesophyll, and an imperfect layer next to the lower epiderm;
large oil-cells in both palisade and sponge, many of those on the lower side being enlarged
elements of the epiderm (84, p. 703). (4) Stomata on lower side only.

Arbutus menziesii (fig. 25). — (2) Bifacial. Palisade tissue of two complete rows, occu-
pying^one-third orjmore of the mesophyll. (3) Lower epiderm minutely papillate. (4)
Stomata on lower 'side only, with small exterior chamber formed by a collar-like ridge.
(5) Tannin very abundant, almost throughout the mesophyll.

Fig. 27. — Dendromecon rigidum: section of leaf.
Fio. 28. — Dendromecon rigidum: stoma. X375.

Climax Chaparral.

Quercus durata (fig. 26).— (2) Imperfectly bifacial. Palisade tissue about three layers
deep, making half the mesophyll; sponge rather loose, but cells palisade-like. (3) Lower
epiderm papillate. (4) Stomata on lower side only. (5) Tannin in lower epiderm and to
some extent in upper; struts of mechanical tissue as in Q. agrifolia.

Berberis pinnata.—(2) Completely bifacial. Palisade tissue two layers deep, occupying
half the mesophyll; sponge fairly typical. (4) Stomata on lower side only. (5) Struts
of mechanical tissue associated with veins.

Dendromecon rigidum (figs. 27, 28).— (2) Almost perfectly isolateral, in correlation with
its vertical placement. Palisade tissue on both sides of leaf, two perfect rows above, two
imperfect ones below, rather loose; sponge central, occupying one-third of the mesophyll,
loose. (3) Both upper and lower epiderm strikingly papillate. (4) Stomata numerous
and large, on both surfaces, sunken in pits to the depth of the thickness of the epiderm.
(5) Struts of mechanical tissue associated with veins.

Heteromeles arbutifolia.—{2) Bifacial. Palisade tissue about three layers deep, making
less than half the mesophyll; sponge abundant, rather typical. (3) Lower epiderm papil-

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 101

late. (4) Stomata on lower side only, slightly sunken, with a small exterior chamber
formed by a collar-like ring. (5) Tannin in bundle-sheath and almost throughout the
mesophyll, abundant. Hypoderm has been reported by Gerard (see Solereder, 84, p. 303),
but I have been unable to find any.

Adenostoma fasciculatum (plates 20b, 21a).— (1) Leaves needle-shaped, nearly terete,
the morphologically upper side flattish, the lower roughly semicircular in cross-section ;
bundles arranged in a rough circle, the main one being opposite the middle of the upper side.
(2) Palisade tissue about three layers in depth, equally developed on all sides, with wide
gaps below the stomata; sponge central, loose, sharply differentiated from the palisade.

29

Fig. 29. — Xylothermia montana: section of leaf. X125.
Fig. 30. — Xylothermia montana: stoma. X375.
Fig. 31. — Rhamnus crocea: section of leaf. X125.

Fig. 32. — Ceanothus cuneatus: section of leaf; black indicates distribution of

tannin. X75.

(3) Cuticle very thick, equally so on all sides. (4) Stomata on all sides, with exterior
chambers of the depth of the cuticle, almost closed at the mouth and containing plugs of
granular material. (5) Veins surrounded by heavy sheaths of large cells which are filled
with tannin. Leaves from seedlings and sprouts are very different. These will be de-
scribed in another section (p. 109).

Primus ilicifolia.—(2) Imperfectly bifacial. Palisade tissue about four layers deep,
making half the mesophyll, with one or two incomplete layers next to the lower epiderm;
sponge scanty. (4) Stomata on lower side only. (5) Abundant tannin in bundle sheath.

Xylothermia montana (figs. 29, 30).— (2) A bifacial leaf of a novel type. Immediately
below the upper epiderm is a single layer of enormous cells elongated perpendicularly to
the surface, and extending from two-thirds to three-fourths of the distance to the lower
epiderm. They resemble palisade tissue both in profile and cross-section, and may perhaps
be regarded as a modified layer of such. They contain dense masses of tannin which com-
pletely fill the cavities, and which frequently come out whole or in pieces in the process

102 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

of sectioning. According to authors quoted by Solereder (84, pp. 259-260), similar sacs
in other members of the family may contain protein and other substances in addition to
the tannin. Between the tapering lower ends of the tannin sacs are found the true pali-
sade cells, which never reach the upper epiderm. Nearly all of the true mesophyll is
palisade-like in character, the only differentiation being moderate looseness and irregularity
near the lower epiderm. (3) Lower epiderm papillate and upper slightly so. (4) Stomata
on lower side only, sunken to the depth of the epiderm, with small exterior chambers.
(5) Tannin sacs: see above.

Rhamnus californica. — (2) Bifacial. Palisade tissue three to four layers deep, occupy-
ing slightly more than half the mesophyll; sponge loose. (3) Some cells of upper epiderm
doubled, and occasionally one with an apparently gelatinized inner wall, as reported by
Herzog (42). (4) Stomata on lower side only, not specialized. (5) Tannin in bundle-
sheath; a rather thick covering of clustered hairs on lower surface.

Rhamnus crocea (fig. 31). — (2) Bifacial. Palisade tissue two to three layers deep, occu-
pying half the mesophyll; sponge very loose. (3) Lower epiderm two to three times as
thick as the upper, composed of cells elongated perpendicularly to the surface of the leaf,
many of them being doubled; appearing like weakly developed water-storage tissue. (4)
Stomata on lower side only, barely sunken. (5) Tannin in bundle-sheath.

33

34 35

Fig. 33. — Ceanothus papillosus: diagrammatic section of leaf. X15.
Fio. 34. — Ceanothus rigidus: section of leaf, showing a single cavity

and its surroundings; general features are like those of

C. cuneatus (fig. 32). X125.
Fio. 35. — Ceanothus sorediatus: section of leaf. X125.

Ceanothus cuneatus (fig. 32).— (1) One of the thickest of the leaves studied. (2) Bi-
facial. Mesophyll very complex (Solereder, p. 886; Gemoll, 32). Palisade tissue about six
layers deep, not continuous horizontally, occupying the areolae between the closely anasto-
mosing veins, the courses of which are prominently marked by strands of tannin cells;
dense, but becoming looser and simulating sponge in center of leaf. Mesophyll extends
to lower surface around the cavities (see below), in the lower half being of rather dense
sponge nature, with a tendency toward palisade-form near the epiderm. (3) Upper epi-
derm partially and irregularly double, the inner cells mostly large. 1 Lower epiderm,

1 Gemoll (32) states that all the cells of the upper epiderm have gelatinized inner walls, and
says nothing of doubling. The interpretation given here seems not to be open to doubt, however.
My material was killed and preserved in formalin-alcohol and cut by the celloidin method, while
Gemoll obtained his from dried herbarium specimens.

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

103

between cavities, with one or two layers of thick-walled hypoderm behind it, the cuticle
papillate. In the areolae between the veins the lower epiderm is deeply invaginated to form
cavities, extending through half the thickness of the leaf and frequently compound, whose
entrances are often smaller than their interior diameters. The epiderm lining these is
very thin, not cuticularized, and bears an abundance of hairs, which fill the cavities with

Fig. 36. — Arctostaphylos hookeri: section of leaf.
Fiq. 37. — Arctostaphylos vestita: section of leaf.

X125.
X125.

Fig. 38. — Eriodictyon californicum: section of leaf. X75.

a dense felt. (4) The stomata are found within the cavities, and are raised above their
immediate surroundings. (5) Tannin is very abundant: in sheath surrounding bundles
and filling very large cells which in the region of the bundles are massed from epiderm to
epiderm, the individual cells being many times larger than the palisades; also in the cells
of the lower epiderm, where it lines the cavities, and in an occasional cell of the upper
layer of the palisade tissue.

104 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

Ceanothus papillosus 1 (fig. 33). —(1) Leaf strikingly revolute. (2) Bifacial. Palisade
tissue two to three layers deep, occupying one-half the mesophyll; sponge rather loose,
and many of the cells palisade-like. (3) Upper epiderm making one-sixth to one-fifth the
thickness of the leaf, composed of large cells filled with tannin; lower epiderm much thin-
ner, undulate, the cells also containing tannin. (4) Tannin in bundle-sheath, collenchyma
of midrib, and epiderm; sparse scattering of single hairs on both sides.

Ceanothus rigid™ (fig. 34). — The description of C. cuneatus will answer for this species
in almost every particular. The slight differences noted are the smaller proportional size
of the entrances to the cavities and the greater elevation of the stomata above their imme-
diate surroundings. Gemoll (32) gives an illustration of the cross-section of the leaf of
C. crassifolius. It is very similar to C. cuneatus and C. rigidus. Probably all the species
of the subgenus Cerastes have leaves of the same general character.

Ceanothus sorediatus (fig. 35).— (2) Bifacial. Palisade tissue two layers deep, occupying
half the mesophyll; sponge rather loose. (3) Epiderm much like that of C. papillosus,
but the upper is even more prominent, making from one-fourth to one-third the thickness
of the leaf. (4) Stomata on lower side only. (5) Tannin abundant in bundle-sheath,
collenchyma of midrib, upper and lower epiderm, and occasional palisade cells; very sparse
hairy covering on lower surface.

Garrya elliptica.—(2) Bifacial. Palisade tissue about three layers deep, occupying half
the mesophyll, the uppermost layer composed of cells that are shorter than the others
and of twice the diameter;* sponge loose. (3) Upper and lower epiderm papillate; upper
cuticle by far the thickest of all those examined. (4) Stomata on lower side only, very
numerous and large, each surrounded by an imposing collar-like ridge. (5) Dense hairy
covering on lower surface; isolated sclerenchymatous cells of various shapes in the meso-
phyll.

Arctostaphylos hookeri (fig. 36).— (2) Almost perfectly isolateral. Mesophyll composed
entirely of palisade tissue— about seven layers— slightly denser on the morphologically
upper side. (3) Cuticle very slightly thinner on the lower side. (4) Stomata on both
surfaces, sunken to the depth of the epiderm, with a small exterior chamber nearly closed
at the mouth. (5) Tannin almost throughout the mesophyll.

Arctostaphylos pumila. — Very similar to the last, but not so perfectly isolateral; palisade
less dense in the lower half; stomata on lower side only.

Arctostaphylos tomentosa. — Essentially like A. pumila. According to Niedenzu (70), the
stomata occur on both surfaces. My material does not confirm this, but such a difference
would not be strange in so variable a species.

Arctostaphylos vestita (fig. 37).— Essentially like A. pumila; cuticle distinctly stratified.

Eriodictyon calif ornicum (fig. 38).— (2) Bifacial. Palisade tissue very sharply differen-
tiated, about four layers in depth, occupying half or a little more than half of the meso-
phyll; sponge loose. (3) Upper epiderm papillate, with an exceedingly thin cuticle; lower
epiderm moderately invaginated between the veins, very thin here, not perceptibly cuticu-
larized, on the ridges opposite the veins resembling the upper epiderm; a very deep invagi-
nation on each side of the very large midrib. (4) Stomata on lower side only, confined
to the invaginated portions, slightly raised above their immediate surroundings. (5) A
dense hairy covering on lower surface, not confined to the furrows, but best developed there.

Two species which belong to the broad-sclerophyll element of the
redwood-forest undergrowth have been studied for the sake of com-
parison. These, with the other members of the group, range far
north of California into the conifer forests of the Puget Sound region.

Gaultheria shallon.—(l) Much like species of Arctostaphylos in general appearance, but
distinctly bifacial. (2) Palisade tissue one to two layers in depth, occupying one-third or
a little more of the mesophyll; sponge rather typical, very loose. (3) Upper cuticle moder-
ately thick, lower thin. (4) Stomata on lower side only, with a small exterior chamber
with constricted opening. (5) Tannin in bundle-sheath, almost throughout mesophyll, and
in upper and lower epiderm.

1 A recent visit to the locality of collection seems to indicate that the material studied is
Ceanothus dentatus rather than C. papillosus. The two species are closely related.

2 Apparently interpreted as hypoderm by Sertorius (see Solereder, 84, p. 433) ; but the cells
contain chloroplasts and are plainly a part of the palisade.

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 105

V actinium ovatum (fig. 39).— (2) Bifacial. Palisade tissue two layers in depth, occu-
pying one-third or less of the mesophyll, its cells very short and rounded; sponge very
loose. (3) Cells of upper epiderm large, with a moderately thick cuticle; lower cuticle
very thin. (4) Stomata on lower side only, very slightly elevated. (5) Tannin in bundle
sheath and in the mesophyll, mainly the sponge-cells.

Table 14. — Summary of structural characters.

Leaf more than 300 microns thick . .
Mesophyll:

Bifacial

Imperfectly bifacial

Isolateral

Entirely palisade

More than 2 layers of palisade .

Sponge central

Water-storage tissue

Epiderm:

Partially double:

Upper

Lower

Lower epiderm invaginated
Papillate :

Upper

Lower

Upper cuticle more than 4 mi-
crons thick

Hypoderm:

Upper side

Lower side

Stomata:

Lower surface only

Both surfaces

Sunken

With exterior chamber

In furrows or cavities

Tannin:

In bundle-sheath

In mesophyll

In epiderm

Mechanical tissue (struts)

Broad-

sclerophyll

forest

(7 species).

Climax

chaparral

(19 species).

12

12

4

3

2

15

2

3
1
3

4
8

13

15
4
6
6
3

9
8
5
3

Total
(26 species).

13

16

7

3

2

18

3
1
3

4
10

14

2
3

22
4
6

7
3

11
9
5
7

Redwood

undergrowth

(2 species).

The principal structural characters of the species described are
summarized in table 14 by communities. From the data given,
we may characterize the average Californian broad-sclerophyll leaf
as follows :

It is moderately small (averaging 2 to 3 cm. in length), simple,
unlobed, elliptic, and in a majority of cases entire and glabrous.
Important groups are toothed, spiny-toothed, revolute, or pubescent
on the lower or on both surfaces. The leaf is thick, averaging
314 microns, while the deciduous species studied average only
127 microns. The mesophyll is most commonly bifacial, though often
imperfectly so, but a few are isolateral. The palisade tissue is

106 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

several layers deep. The epiderm is nearly always single, and often
papillate. The cuticle is very thick, the upper averaging 5.69
microns as against 1.58 microns in the deciduous species. In the
case of the lower the difference is still greater: 4.43 microns and
0.91 microns. The stomata in a large majority are on the lower
surface only. Many special features occur in their structure and
distribution that are effective in decreasing water-loss. Tannin is
abundant and widespread in bundle-sheath, mesophyll, and epiderm.
Divergences from Schimper's generalizations, quoted at the
beginning of the chapter, are unimportant. The average Californian
leaf is smaller than his statement implies; it tends toward the oval
rather than the lanceolate in form, and vertical placement is less
pronounced. Many special features are here described which do
not appear in his account.

Comparing the two communities, we find that the broad-sclerophyll
forest has uniformly much larger leaves, which are less often entire.

The climax chaparral leaf averages
50 per cent thicker than the forest
leaf; it includes all the isolateral
and the majority of the imper-
fectly bifacial leaves, and its pali-
sade tissue is more prominent,
averaging 4 to 5 layers and 66 per
cent of the mesophyll as against
2 to 3 layers and 43 per cent in
the forest species. The chaparral
leaves have a much thicker cuticle
than those of the forest — nearly
100 per cent greater on the upper
side and more than 150 per cent
greater on the lower. The species with stomata on both surfaces,
which are those with isolateral vertical leaves, are all of the chaparral.
Special features protective against water-loss are almost entirely
confined to the chaparral.

We see thus that the broad-sclerophylls, as compared with decid-
uous species, have a wealth of features of the kind that have been
assumed to be a response to the moisture conditions of the habitat,
and which are certainly effective in decreasing water-loss. Certain
of the characteristics, such as spiny teeth, papillate epiderm, and
presence of tannin, have no obvious relation to this or to any other
habitat factor, either as to cause and effect or advantage. We see
further that the chaparral species are in every way more fully provided
with transpiration-decreasing features than are the forest species.
The group from the redwood undergrowth is rather puzzling. In
thickness of leaf the two species studied are surpassed by but three

Fig. 39. — Vaccinium ovatum: section of leaf.
X125.

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 107

species of the chaparral; in thickness of cuticle they are superior to
all the species of the forest and to many of the chaparral. The
mesophyll, however, is decidedly suggestive of mesophytic influences.
The habitat of these plants is far more mesophytic than that of the
other broad-sclerophylls, and the xeromorphic construction of their
leaves is a warning against the determination of the essential ecolog-
ical nature of a plant by one structure alone. The commonness of
this type of leaf in the family Ericaceae, apparently regardless of
habitat, is too well known to require comment.

EFFECTS OF ENVIRONMENTAL CONDITIONS UPON LEAF STRUCTURE.

It is commonly assumed that moisture and light influence power-
fully the structure of leaves, especially the thickness and character
of the mesophyll and epiderm. Many have essayed to separate
the effects of the two factors, not always with success. While it is
impossible in such a field study as the present to make a close analysis
of cause and effect, any correlation of structural differences with
accurately measured habitat factors is of value, and therefore the
following brief study is presented.

Three stations in the series at Jasper Ridge (p. 33) were selected,
Nos. 4, 3, and 7 representing respectively the Adenostoma, the
Arctostaphylos, and the Quercus agrifolia-Arbuius communities.
A summary of the moisture and light conditions of each station is
presented in table 15. The figures are derived from the data given
in an earlier section (p. 42). Soil-moisture is expressed in per cent
of dry-soil weight, and the figure is an average of a series of six
monthly determinations at two depths (10 cm. and 30 to 50 cm.)
during the dry season from June 15 to October 31, 1913. Evapora-
tion is given in cubic centimeters of daily loss from a standard
atmometer, determined weekly, and averaged for practically the same
period as the soil-moisture readings. Light is expressed as a decimal
of the full illumination. The readings were made as close together
as possible in time, at midday of August 17, 1917. The species
chosen for study were the three most prominent chaparral shrubs of
the vicinity, Adenostoma fasciculatum, Arctostaphylos tomentosa, and
Quercus durata. The natural habitat of these species is of the character
of stations 4 and 3, but they grow occasionally as interlopers in the
forest, which is represented by station 7, and thus give opportunity
for comparison. The material was collected from widely separated
plants within a given association, preserved in formalin-alcohol and
cut in paraffin. In making the measurements, 10 separate leaves
were used and 5 measurements made upon each leaf. Every figure
given is therefore an average of 50 separate measurements.

The habitat analysis shows that the stations, in the order as given,
are progressively more favorable as regards moisture conditions,
both in water-content and in evaporation, and progressively more

108 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

deficient in light. Stations 3 and 4, however, are rather similar in
every factor, while station 7 is quite different. It is not surprising,
therefore, to find striking structural divergence between stations
3 and 4 on the one hand and station 7 on the other, and far less
consistent differences between stations 3 and 4. The leaf is thinnest

Table 15. — Effect of habitat upon leaf structure.

Station 4.

Adenostoma

chaparral.

Station 3.

Arctostaphylos

chaparral.

Station 7.

Quercus agrifolia-

Arbutus forest.

Habitat:
Light:

Sun

0.75
0.21
3.24
30.1

541
10.10
10.26

371
9.77
9.36

233

3.38
4.05

0.69
0.04
3.38
28.4

574
10.65
11.00

366
9.06
14.30

198
2.77
3.23

0.56
0.07
3.81
22.7

475
11.00
4.64

250
6.94
7.67

193

2.55
2.66

Leaf -structure:

Adenostoma f asciculatum :

Upper cuticle, microns

Lower cuticle, microns

Arctostaphylos tomentosa:

Upper cuticle, microns

Lower cuticle, microns

Quercus durata:

Upper cuticle, microns

Lower cuticle, microns

in station 7, and the same is true of the lower cuticle, the differences
in the latter case being especially noticeable. In the case of the
upper cuticle the differences in Arctostaphylos and Quercus are less,
and in Adenostoma there is reversal, the mesophytic station 7 showing
a cuticle slightly thicker than the others. There are also differences
apparent in the mesophyll, especially in Arctostaphylos. In this
species the sections from stations 3 and 4 are barely distinguishable,
but these two differ strikingly from station 7. Such differences
can not be expressed numerically, but are very evident in drawings
(figs. 40 and 41). In the leaf from station 4, the dense palisade extends
from epiderm to epiderm, being slightly looser below, with here and
there a tendency toward sponge character. In that from station 7
the palisade is but two layers deep, and there is a distinct region of
sponge tissue. Quercus and Adenostoma from the three stations show
far less striking differences.

We conclude, therefore, that increased thickness of leaf and of
cuticle, increased development of palisade tissue, and decrease of
sponge coincide with decrease of moisture and increase of light.
Further than this we can not certainly go. The natural assumption
is that moisture is the controlling factor rather than light, which
may have an indirect effect.

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 109

Evidence that this assumption is well founded may be obtained
from that unusually interesting plant Adenostoma fasciculatum, in
the remarkable type of leaf that is found upon seedlings and stump
sprouts and to a slight degree on mature plants growing in meso-
phytic situations. Prints of these, with the ordinary type for com-
parison, are given in plate 21b. The differences in structure are no
less striking than in form (plate 20, b, c, d). The sprout or seedling

Fig. 40. — Arctostaphylos tomentosa: section of leaf from a xerophytic habitat (station 4,

Jasper Ridge). X125.
Fig. 41. — Ardostaphylos tomentosa: section of leaf from a mesophytic habitat (station 7,

Jasper Ridge). X125.
Fig. 42. — Adenostoma fasciculatum: stoma of normal xerophytic specimen. X375.
Fig. 43. — Adenostoma fasciculatum: stoma of leaf from stump sprout. X375.

leaf is a flat isolateral structure, averaging 300 microns in thickness,
with the veins in a single plane after the manner of ordinary leaves.
The mesophyll is loose, especially the central sponge, and both
palisade and sponge cells are short and rounded. The epidermal
cells are very large, thin-walled, and turgid. The cuticle is very
thin — 1.5 microns as contrasted with 9.73 microns in the ordinary
leaf of the species. The stomata occur on both surfaces. They still
possess the exterior chamber, but this is very shallow and open-
mouthed (figs. 42, 43).

We may explain these remarkable structural changes — broadening
of the leaf to form a flat blade, loosening of the mesophyll, loss of
character in the palisade, reduction of cuticle — after a fashion by

110 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

stating that the abnormal leaf type is an ancestral character to which
the plant reverts in its juvenile stage. This in itself is begging the
question, and doubt is cast upon its adequacy by the fact that the
flat leaf sometimes appears in mature individuals in mesophytic
situations (plate 21a, 6). Light can not be operative here, since it
is at maximum intensity where sprouts and seedlings are found, and
observed changes due to increased intensity are invariably in the
opposite direction. Moisture as the governing factor seems entirely
adequate. New stump sprouts, even in the midst of the dry season,
have an exceptional advantage as to water, since they may draw
upon an absorbing system which formerly supplied a complete
mature shrub. The large size of leaves upon stump sprouts and
their mesophytic structure are well-known phenomena, and are
characteristic of other chaparral species besides Adenostoma. As
to seedlings, germination and early growth take place at a time of
year when soil-moisture is abundant and evaporation low, and the
plant supplements this advantage by immediately sending down a
taproot to a considerable depth. The occasional occurrence of the
flat leaf upon mature individuals in mesophytic situations is the
final argument. It may well be that the mesophytic leaf with its
distinctive form and structure points back to a mesophytic ancestry
for the species, which would help to bring it into harmony with the
majority of the members of its family, among which it is decidedly
an anomaly. Finally, since moisture seems adequate as a cause of
variation in the form and thickness of the leaf of Adenostoma and
in the character of its mesophyll and cuticle, it is reasonable to assume
that it is the controlling factor in the other species where the same
differences are present but less striking.

COMPARATIVE TRANSPIRATION-RATE.

In the summer of 1917 I made a brief study to determine the actual
transpiration-rate of two or three of the important broad-sclerophyll
species. The results were somewhat puzzling, though consistent
enough among themselves, and it is therefore the better part of valor
to withhold them in the main for further investigation. One or two
points were clearly brought out, however, which are worth stating
here.

The study was made by the simple potometer method — the placing
of branches in bottles of water, which were weighed at the beginning
and end of the experiment. The mouths of the bottles were plugged
with cotton, which was kept perfectly dry. One or two which became
wet during the experiment were discarded. The species used were
Adenostoma fasciculatum , Arctostaphylos tomentosa, and Arbutus
menziesii. Ten shoots of each of the first two were cut in station 10
{Adenostoma consociation; see p. 41 for environment), placed in

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. Ill

bottles and weighed. Five of each species were left in station 10
and the rest were placed in station 7 (Quercus agrifolia- Arbutus associa-
tion, p. 38). In station 7, 10 branches each of the same species
were cut, and in addition 10 of Arbutus menzicsii, and half of each
set were left in station 7 and the others taken to station 10. The
experiment was allowed to run for about 48 hours, from August 24
to August 26. The branches were pressed between driers, and the
areas of the leaves of Arctostaphylos and Arbutus were later deter-
mined with a planimeter. In the case of Adenostoma, with its
terete leaves, a different and less exact method was necessary. Each
branch was stripped of its leaves and these counted. Then 20 were
selected at random, placed in boiling water to restore them approxi-
mately to their original size, and measured accurately as to length and
diameter. The area was computed as a cylindrical surface, the
average of the 20 taken, and this applied to the total number of
leaves on the shoot. The total losses recorded for each shoot were
reduced to the amount per square decimeter of surface. So much of
the results as are worth giving at present are seen in tables 16,
17, and 18.

Table 16. — Loss per square decimeter, station 10.

c. c.

Adenostoma (average of 5 shoots) 5 32

Arctostaphylos (average of 5 shoots) 6 . 68

Arctostaphylos (both surfaces considered) 3.34

It is shown that when the stoma-bearing surface alone is con-
sidered, the loss per unit is less in Adenostoma; but if both surfaces
of the leaf are considered in Arctostaphylos, as we must fairly do,
the relation is reversed. It would seem, therefore, that the Adeno-
stoma leaf is the more effective in reducing water-loss so far as the
actual stoma-bearing surface is concerned, but that Arctostaphylos
compensates for this by the entire absence of stomata (at least in
the plants studied; see p. 104) over half its leaf-surface. The more
perfect xerophytism of Adenostoma can not be explained, therefore,
upon the basis of cuticular and stomatal regulation; and this is
confirmed by the descriptions presented in the last section, in which
these structures are seen to be very similar in the two species. We
may fall back upon the hypothesis that the total leaf-area is less
in Adenostoma than in Arctostaphylos, plants of equal size being con-
sidered; individuals of the two species are apt to be so when growing
together. The general appearance would indicate this, especially
the small size of the leaves of Adenostoma, and confirmation is ob-
tained from the fact that the average leaf-area of the shoots used in
the experiment (both surfaces considered in Arctostaphylos) was
75 per cent greater in Arctostaphylos. Of course, this is merely a
rough approximation, but since care was taken to select branches
as nearly alike as possible, it doubtless indicates the truth.

112 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

For comparison with Arbutus it is necessary to use the shoots of all
species that were cut in station 7 and left there (table 17).

Table 17. — Loss per square decimeter, station 7.

e. e.

Adenostoma (average of 5 shoots) 2 . 77

Arctostaphylos (average of 5 shoots) 3 96

Arctostaphylos (both surfaces considered) 1 . 98

Arbutus (average of 5 shoots) 3 . 69

Arbutus (both surfaces considered) 1-85

The two shrubs show the same relations here as in station 10.
Arbutus seems to be slightly more effective in reducing water-loss
than Arctostaphylos, and therefore more effective than Adenostoma
per unit total surface. However, the average total leaf-surface
per shoot was 38 per cent greater than Arctostaphylos and thus far
greater than Adenostoma, indicating that even for a plant of equal
size the total leaf-surface would be greater; and there is the further
fact of the tree stature of Arbutus. The conclusion seems to be
that it is not so much differences in cuticular or stomatal effectiveness
that make the difference in drought resistance as it is the total leaf-
area exposed in plants of approximately equal size, the separate
elements in this character being number and size of leaves.

One other point may be brought out. Table 18 exhibits the
difference in evaporation-rate which appears in the same species
growing in the two habitats, due fundamentally to differences in
the environmental conditions.

Table 18. — Loss per square decimeter, station 7.

c. c

Adenostoma (station 7) 2 . 77

Adenostoma (station 10) 5 . 32

Arctostaphylos (station 7) 1-98

Arctostaphylos (station 10) 3 .34

In Arctostaphylos the rate in the xerophytic station is 70 per cent
greater than in the other, while in Adenostoma the difference is
nearly 100 per cent. The study of the transpiration of these plants
is one of the most promising fields for further investigation.

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 113

APPENDIX.

ANNOTATED LIST OF BROAD-SCLEROPHYLLS AND ACCOMPANYING

SPECIES.

In listing the species which make up the broad-sclerophyll vegeta-
tion of California, a separation into groups will be convenient. The
first great division is into dominants and secondary species; both
classes may conveniently be subdivided.

THE DOMINANTS.

Under this head I would include all the true forest and brush-
forming species — those which give to the type its characteristic
physiognomy. Although the majority are strikingly similar in
ecological character, they present great floristic diversity. The
first two lists include 74 broad-sclerophylls, representing 23 genera
and 13 families. In all lists species restricted to the Calif ornian
islands have been excluded. Another group of broad-sclerophylls,
mentioned in the last chapter, comprises the half-dozen species
which form an important part of the undergrowth of the northwestern
conifer forests. These are Berberis nervosa Pursh, Pachystima
myrsinites Raf., Rhododendron californicum Hook., Gaultheria shallon
Pursh, and V actinium ovatum Pursh.

The following list includes all the species that are of importance
in the composition of the broad-sclerophyll forest. They are few
in number, and most are of wide range within the Californian region.
Some are themselves not broad-sclerophylls, but are deciduous;
these have been grouped together at the end of the list. A number
of the species, too, when growing in relatively unfavorable situations,
take the shrub form and in some cases are more important thus than
as trees. In a few instances the shrubby form has been separated
by taxonomists as a variety.

List I. — Arborescent Species.
EVERGREEN.

1. Myrica californica Chamisso. Wax myrtle. Immediate coastal region from Puget
Sound to Los Angeles County. Of subordinate importance as a forest tree.

2. Castanopsis chrysophylla (Hook.) A. DC. Golden chinquapin. As a forest tree it
occurs in western Oregon and in the mountains of northwestern California southward to
Mendocino County. In the latter region it is a member of the mixed broad-sclerophyll-
conifer forest. It is also frequently associated with the redwood. Var. minor A. DC. is a
shrub closely resembling C. sempervirens (see No. 17). It is characteristic of the Coast
Ranges, extending southward to Monterey, while C. semperrirens belongs rather to the
Sierras. It is a frequent member of the conifer forest chaparral in the north Coast Ranges,
and appears also in the climax chaparral.

3. Quercus agrifolia Nee. Coast live oak. Outer Coast Ranges and valleys near the
coast, from Mendocino County to Mount San Pedro Martir in Lower California. The
most important live oak, and probably the most important of the genus, within its range.
Covers hillsides, especially north-facing slopes, with dense forest; also commonly in open
park-like growth in valleys of the coast range; in such places attaining its largest size, and
frequently associated with Q. lobata. Almost invariably tree-like in form, the only notable
exceptions being the bushy contorted specimens in exposed situations near the shore.

114 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

4. Quercus chrysolepis Liebm. Mountain live oak. In southwestern Oregon and
throughout the California mountains west of the high Sierras and the desert, except the
lowest foothills; south to Mount San Pedro Martir in Lower California. The most widely
distributed and most important of the live oaks of California, being the dominant member
or one of two or three dominants in the broad-sclerophyll forest wherever it occurs. Very
frequent in chaparral form, sometimes in rather tall pure growth on north slopes, forming

1 dwarf for est."

5. Quercus engelmanni Greene. Engelmann oak. Of limited range; in southern Cali-
fornia from Los Angeles County to the Mexican boundary, occupying a belt 80 km. wide,
distant 25 to 30 km. from the coast. Important within its limited range. Barely ever-
green.

6. Quercus wislizeni A. DC. Interior live oak. From Mount Shasta southward through
the Sierra foothills and the Coast Ranges to Mount San Pedro Martir in Lower California;
rarely, though occasionally, near the coast. Much less important as a forest tree than
Q. chrysolepis, and characteristic of less mesophytic situations; occurs frequently in open
growth with Q. douglasii on the lowest foothills. The shrubby form, var. frutescens Engelm.,
is very common in the climax chaparral.

7. Pasania densiflora Oerst. Tan-bark oak. Southwestern Oregon to the Santa
Ynez Mountains in the Coast Ranges, and to Mariposa County in the Sierras. Character-
istic of the mixed forest of the interior mountains of northern California; also commonly
associated with the redwoods. The shrubby form, var. echinoides Jepson, is a member of
the conifer-forest chaparral in northern California.

8. Umbellularia calif ornica (H. and A.) Nutt. California laurel. Southwestern Ore-
gon, southward through the Coast Ranges and Sierras to southern California. Character-
istically a tree of the mesophytic forests, especially addicted to steep north slopes, ravines,
and stream-banks. Sometimes shrubby, forming an unimportant part of the chaparral in
the more mesophytic situations, especially in the extreme southern portion of its range.

9. Arbutus menziesii Pursh. Madrono. Southwestern British Columbia; through the
coast region of Washington and Oregon; mountains of northern California; southward in
the Sierras to Tuolumne County, and in the Coast Ranges to Los Angeles County. An ex-
ceedingly important tree of the broad-sclerophyll forest. Largest in northwestern Cali-
fornia, where it accompanies the redwoods; in the central Coast Ranges second in impor-
tance to the live oaks, with them forming the bulk of the north-slope forests; decreasing
southward in size and abundance, both in Coast Ranges and Sierras.

DECIDUOUS.

10. Quercus douglasii H. and A. Blue oak. Low foothills and valleys from Mendocino
County to the upper Sacramento Valley, southward to the Liebre Mountains and the San
Fernando Valley. Usually in open growth on the lowest foothills; sometimes pure, some-
times with Pinus sabiniana or Quercus wislizeni. Rarely associated with the broad-
sclerophyll trees; the most xerophytic of the California oaks. Its relation to the chaparral
is treated on page 79.

11. Quercus garryana Dougl. Garry oak. North Pacific States, extending southward
in California along the Coast Ranges; abundant as far as Trinity County; occasional to
the Santa Cruz Mountains. Not very important in California as a forest tree. In the
northwestern portion, however, it forms dense pure thickets 2 or 3 meters high in the
forest region.

12. Quercus kelloggii Newb. California black oak. Corresponding in range rather
closely to Quercus chrysolepis, and a common companion to it in the forest. Occasionally,
in shrub form, it makes a part of the conifer-forest chaparral.

13. Quercus lobata Nee. Valley oak. Valleys and foothills from Shasta County to
Los Angeles County. Included here because it frequently forms park-like mixed forest
with Q. agrifolia.

14. Acer macrophyllum Pursh. Broadleaf maple. Coast Ranges and Sierras, south to
the San Bernardino Mountains. A frequent companion to the broad-sclerophyll trees in
the more mesophytic localities.

15. ^sculus calif ornica Nutt. California buckeye. Coast Ranges from Mendocino
County to San Luis Obispo County; Sierra foothills. A common inhabitant of north-
facing slopes in the foothills, growing with the live oaks; sometimes forming dense pure
thickets.

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 115

The species of the next list are the shrubs that make up the chapar-
ral, which in extent and in number of species and of individuals
is far more important than the broad-sclerophyll forest. A further
point of contrast is the far greater number of species in the chaparral
which have a decidedly restricted range. To complete the picture,
the following species from List I should be added, since in shrub form
they have a more or less important place in the chaparral :

Castanopsis chrysophylla minor. Quercus wislizeni frutescena.

Quercus chrysolepis. Pasania densiflora echinoides.

Quercus garryana.

Considering the region as a whole, 4 genera, representing as many
families, are of paramount importance in the chaparral. These
are Quercus, Adenostoma, Ceanothus, and Arctostaphylos. They
are widely different in the number of species representing them.
Adenostoma, with the" most important single species, includes 2;
Quercus has 6 that are strictly chaparral forms; while in Arctosta-
phylos and Ceanothus I have listed 18 and 25, respectively. The
treatment of the last two genera is of necessity unsatisfactory, since
there is disagreement among taxonomists upon the fundamental
questions of specific limits and relationships. In both genera there
are groups of forms which are regarded by some authors as constitut-
ing single variable species. Within a group there are several forms
which may be species, varieties, or merely ecological variants. ^ It
seems better for the present purpose, in doubtful cases, to consider
the larger unit as the species, rather than to attempt to maintain a
number of so-called species, sometimes poorly defined, of uncertain
range, often known only from the type locality. It is not denied
that these variant forms are of the greatest ecological interest and
importance. At the present time, however, our knowledge of them
is in every way insufficient to make adequate discussion possible.
In the treatment of each genus I have, so far as possible, followed
the author who has given the most recent and complete taxonomic
account of the group concerned. These are Trelease (89) in Ceano-
thus and Abrams (2) in Arctostaphylos. In both cases I have omitted
a number of species of uncertain validity or range.

In the list below I have attempted to indicate as accurately as
has been possible to determine, from personal observation, literature,
and herbaria, the range and degree of importance of each species of
the chaparral. In the case of the most important one, Adenostoma
fasciculatum, the range has been plotted on a map (plate 3). This is
fairly accurate, since it is based on a careful and widely extended
field study, supplemented by examination of literature and herbarium
records, and by notes from a number of persons possessing a wide
acquaintance with the vegetation of California. These sources
together yielded a total of about 250 known localities.

116 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

List II. — Obligate Shrubs.
BETULACEjE.

16. Corylus rostrata Ait. var. californica A. DC. Hazelnut. Central and northern
Coast Ranges and Sierra Nevada. A rather unimportant constituent of the conifer-forest
chaparral; occurs more commonly as undergrowth in mesophytic forest. Deciduous. 1

17. Castanopsis sempervirens (Kellogg) Dudley. Golden chinquapin. From the south-
ern Cascades along the Sierras, southward to Mount San Jacinto. Characteristic of the
conifer-forest chaparral of the higher altitudes.

18. Quercus breweri Engelm. Brewer oak. Inner Coast Ranges from Mount Yolo
Bolly north; Sierra Nevada to the Kern River region. Forms extensive thickets at middle
altitudes. A member of the conifer-forest chaparral. Deciduous.

19. Quercus dumosa Nutt. Scrub oak. From the region of Mount Shasta to Lower
California (lat. 31°); on the Coast Ranges and Sierra Nevada. Most abundant in the
coastal region of southern California. A very important member of the climax chaparral,
ranking with certain species of Arctostaphylos and Ceanothus, next to Adenostoma. Barely
evergreen.

20. Quercus durata Jepson. Scrub oak. Closely related to the last. Range not satis-
factorily determined, but apparently covering the middle Coast Ranges from the Santa
Lucia Mountains north to Mendocino County and the Napa Range, and probably to the
Trinity Mountains. Reported from Stanislaus National Forest in the Sierra Nevada.
Similar in status to the last.

21. Quercus sadleriana R. Br. Confined to the mountains of southwestern Oregon and
northwestern California. Forms pure thickets, probably successional. Barely evergreen.

22. Quercus vaccinifolia Kellogg. Huckleberry oak. Inner north Coast Ranges from
Mount Shasta to Trinity County; throughout the higher forest region of the Sierra Nevada.
A member of the conifer-forest chaparral; also occurring as undergrowth in the subalpine

TOT*PSts

BERBERIDACE^l.

23. Berberis pinnata Lag. California barberry. Central Coast Ranges. Sometimes
in the chaparral; occasionally in the oak forests. Unimportant.

PAPAVERACEjE.

24. Dendromecon rigidum Benth. Bush poppy. Of general distribution; from Shasta
National Forest southward through Coast Ranges and Sierras, reaching Mount San Pedro
Martir in Lower California; records scarce north of Sonoma and Napa Counties. A
widely distributed and characteristic member of the climax chaparral, though nowhere
abundant; its bright yellow flowers contribute a pleasing touch of color here and there.

ROSACEA.

25. Heteromeles arbutifolia (Poir.) Roem. Christmas berry; California holly; tollon
(pronounced "toyon"). Coast Ranges and Sierra Nevada, and through southern Califor-
nia and Lower California to the Cape region. A characteristic member of the climax
chaparral; not often abundant, but seemingly so because of its fine clusters of red berries
and because it usually overtops its neighbors. Sometimes attains tree stature, but retains
its shrub form. More tolerant of shade than most of its chaparral companions; therefore
frequently found in the oak forest.

26. Arnelanchier alnifolia Nutt. Service berry. Has the widest distribution of all the
species occurring in the chaparral, ranging from Alaska to Mount San Pedro Martir in
Lower California, east to Michigan, Nebraska, Colorado, and New Mexico; in California
occurring in the Coast Ranges from San Francisco Bay northward, in the Sierras and the
higher mountains of southern California. A rather frequent member of the conifer-forest
chaparral. Deciduous.

27. Cercocarpus betuloefolius Nutt. Birch-leaf mahogany. Throughout the foothills
and lower mountains of California. A common and frequently abundant member of the
climax chaparral, especially in the more mesophytic situations. Often gregarious on north
slopes, the groups in seed time showing from a distance as white patches.

1 Unless otherwise'notcd, species are evergreen in leaf habit.

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 117

28. Cercocarpus ledifolius Nutt. Mountain mahogany. Ranging widely over the west-
ern States, but not important as a member of the chaparral. Occurs sparingly in the
chaparral of the San Bernardino Mountains and other southern California ranges, and in
the Sierras, but mainly on the east slope.

29. Adenostoma fasciculatum H. and A. Chamise; grease-wood (the latter name,
though frequent, is unfortunate, as it has been commonly applied to other shrubs, totally
unrelated, of the western United States). Range shown in detail on map (plate 3). Its
occurrence near Hershey in the Sacramento Valley, and other facts, seem to indicate that
it formerly had a more widely extended range (see p. 79). The most abundant species and
most important in every way of the shrubs of the climax chaparral. Usually in pure
growth or nearly so over extensive areas, occupying the less mesophytic situations, which
greatly predominate in the chaparral region. Pure growth is commonly known as "cha-
misal." Easily recognized at a distance, when growing in mass, by its characteristic
color-tone, which varies according to the time of year. Its general shade is gray-green,
but in June this is whitened by the profusion of spiraea-like flower panicles, and through
the remainder of the summer and autumn turned to rich brown by the equally abundant
clusters of withered flowers and achenes.

30. Adenostoma sparsifolium Torr. Yerba del Pasmo. San Jacinto and Santa Monica
Mountains to Lower California; also in Santa Ynez Mountains. An interesting and hand-
some species of limited range and abundance; occurs in the same sort of situations as the
last.

31. Prunua demissa (Nutt.) Walp. Western choke cherry. Rocky Mountains to the
Pacific States and British Columbia. In California, a widely distributed but unimportant
member of the conifer-forest chaparral; commoner in stream-bank thickets. Deciduous.

32. Prunus emarginata (Dougl.) Walp. Bitter cherry. Range similar to last; not so
far eastward, but extending south to Mount San Pedro Martir in Lower California. In
California an important member of the conifer-forest chaparral, frequently forming exten-
sive thickets. Deciduous.

33. Prunus ilicifolia (Nutt.) Walp. Holly-leaf cherry; islay. South Coast Ranges,
from the region of San Francisco Bay to Lower California. A frequent member of the
climax chaparral; when shrubby closely resembling Rhamnus crocea. In more mesophytic
situations it often attains small tree size, with a trunk 3 dm. thick, in such cases resembling
Quercus agrifolia so closely in habit and leaf-form as to be distinguishable with difficulty.

34. Prunus subcordata Benth. Western plum. Southern Oregon southward, in the
Coast Ranges to the Mount Hamilton Range, in the Sierras to the Kern River region. An
unimportant member of the conifer-forest chaparral. Deciduous.

CESALPINACEiE.

35. Cercis occidentalis Torr. Redbud. Foothills of Sierras and Coast Ranges south-
ward to San Diego County, eastward to western Texas. Frequently a stream-bank shrub;
occasionally a member of the chaparral; in the Cuyamaca Mountains seen in pure thicket
growth on a northeast slope. Deciduous.

LEGUMINOSiE.

36. Xylothermia montana (Nutt.) Greene [Pickeringia montana Nutt.]. Chaparral pea.
Coast Ranges from Mendocino and Lake Counties southward; Sierra Nevada, from Mari-
posa County southward; var. tomentosa Abrams in mountains of southern California and
Lower California. A widely distributed but not abundant member of the climax chaparral.

RUTACE^E.

37. Cneoridium dumosum (Nutt.) Hook. f. San Diego County and northern Lower
California. Climax chaparral.

ANACARDIACEjE.

38. Rhus integrifolia (Nutt.) B. and H. Mahogany sumach. Santa Barbara to Mount
San Pedro Martir and Magdalena Bay, extending eastward to the desert slopes of the
Cuyamaca Mountains.

39. Rhus laurina Nutt, Laurel-leaf sumach. Santa Ynez Mountains to the San Ga-
briel Range, southward to northwestern Lower California, mainly near the coast. Usually
not abundant, but conspicuous by reason of its light-green foliage.

40. Rhus ovata Wats. Ovate-leaved sumach. Santa Ynez Mountains to San Bernar-
dino and San Jacinto Ranges, south to Mount San Pedro Martir in Lower California.

118 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

ACERACE.E.

41. Acer glabrum Torr. Dwarf maple. Southeastern Alaska to Montana, Colorado,
western Nebraska, New Mexico; west to California, where it ranges from the Oregon line
southward to northern Trinity County, throughout the middle altitudes of the Sierras, and
in the San Bernardino and San Jacinto Ranges. A mesophytic shrub, sometimes forming
an unimportant part of the conifer-forest chaparral. Deciduous.

RHAMNACE^l.

42. Rhamnus californica Esch. (including var. rubra Trelease and var. tomentella Brew,
and Wats.). Coffeeberry. Coast Ranges and Sierra Nevada, south to Mount San Pedro
Martir in Lower California. A frequent member of the climax chaparral, occurring also
in more mesophytic situations such as the oak forest. Var. tomentella extends eastward into
Arizona and New Mexico; var. rubra is deciduous, belonging rather to the conifer-forest
chaparral.

43. Rhamnus crocea Nutt. (including R. ilicifolia Kellogg). Evergreen buckthorn; red-
berry. Siskiyou County southward in the Coast Ranges and Sierra Nevada, extending to
Mount San Pedro Martir in Lower California; eastward to Providence Mountains, Cali-
fornia, and into Mexico and Arizona. A frequent member of the climax chaparral.

44. Ceanoihus cordulatus Kellogg. Snowbrush. Forested northern Coast Ranges and
throughout the higher Sierra Nevada forest regions; higher mountains of southern Cali-
fornia and Lower California to Mount San Pedro Martir; mountains of Nevada according
to Mrs. Brandegee (10). One of the most important constituents of the conifer-forest
chaparral, covering extensive areas pure or mixed with other species. As is natural, con-
sidering its successional status, it is also found abundantly as undergrowth in the coniferous

45. Ceanothus crassifolius Torr. Santa Ynez Mountains to the San Gabriel and San
Bernardino Ranges, south to northern Lower California. An important constituent in its
range of the climax chaparral, conspicuous because of its grayish-green foliage.

46. Ceanothus cuneatus Nutt. Wedge-leaf ceanothus. Omnipresent in the lower alti-
tudes of the California mountains; extending northward into southern Oregon and prob-
ably southward into Lower California. One of the three or four most abundant species
in the climax chaparral region. It comes up in great numbers after fire, and its presence
in abundance usually indicates recent disturbance.

47. Ceanothus dentatus T. and G. Outer Coast Ranges from the Santa Cruz Mountains
to Santa Barbara County.

48. Ceanothus divaricatus Nutt. (including var. eglandulosus Torr.). Monterey County
south to the San Bernardino Range, Cuyamaca Mountains, and Lower California; also
in the central and southern Sierras. An important shrub in the climax chaparral of the
southern California mountains.

49. Ceanothus diversifolius Kellogg. Central Sierras, in the yellow pine belt. A creep-
ing shrub, growing beneath the pines, not forming brush.

50. Ceanothus foliosus Parry. Coast Ranges north of San Francisco Bay, to the red-
wood region of Mendocino County.

51. Ceanothus hirsutus Nutt. (including C. oliganthus Nutt.). Range uncertain; south-
ern Coast Ranges; perhaps southward into Lower California.

52. Ceanothus incanus T. and G. Coast Ranges from Humboldt County to the Santa
Cruz Mountains, mainly in the redwood region.

53. Ceanothus integerrimus H. and A. Deerbrush. Mount Shasta region, south in the
Coast Ranges to the Santa Cruz Mountains; lower yellow pine belt of Sierra Nevada; var.
puberulus (Greene) Abrams, in the mountains of southern California. According to Mrs.
Brandegee (10, p. 184), the species ranges southeastward to southern Arizona. An impor-
tant species in chaparral and as undergrowth in coniferous forest. Deciduous.

54. Ceanothus megacarpus Nutt. Near the coast, Santa Ynez to Santa Ana Mountains.

55. Ceanothus Palmeri Trelease. Ventura County to the Cuyamaca Mountains and
perhaps farther south.

56. Ceanothus papillosum T. and G. Of limited range; Santa Cruz Mountains to Santa
Lucia Mountains. Grows in the Monterey region with an interesting group of endemics
and species of restricted range, including C. rigidus, Arctostaphylos hookeri, A. pumila, and

A fjPstitd

57. Ceanothus parryi Trelease. Coast Ranges; Napa and Solano Counties northward
to the redwood region of western Mendocino and Humboldt Counties. Deciduous.

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 119

58. Ceanothus parvifolius (Wats.) Trelease. Central and southern Sierras; conifer-forest
chaparral. Deciduous.

59. Ceanothus pinetorum Coville (including C. jepsonii Greene). Coast Ranges; Lake
County to Mount Tamalpais; also in the southern Sierras (Tulare County).

60. Ceanothus prostratus Benth. Squaw carpet; mahala mats. North Coast Ranges,
Mount Shasta to the San Francisco Bay region; pine belt of the Sierra Nevada. Habit
and habitat of C. diver sifolius, except that toward its southern limit in the Coast Ranges it
tends to become erect (var. divergent K. Brandegee) .

61. Ceanothus rigidus Nutt. Limited in range; Marin County to Monterey. (See No.
56.)

62. Ceanothus sanguineus Pursh. A northern species reaching the Siskiyou Mountains
in northern California; conifer-forest chaparral. Deciduous.

63. Ceanothus sorediatus H. and A. Coast Ranges; Napa and Solano Counties to north-
ern Santa Barbara County. Somewhat like C. cuneatus in its habits, but preferring more
mesophytic situations; coming up thickly like C. thyrsiflorus after fire in the redwood
forest.

64. Ceanothus spinosus Nutt. Coast region of southern California from Santa Barbara
County to Orange County. One of the largest members of the genus, sometimes arbor-
escent. Partly evergreen.

65. Ceanothus thyrsiflorus Esch. Blue-blossom; California lilac. An abundant spe-
cies of the redwood region from the northern boundary of the State to the Santa Lucia
Mountains; temporary, forming dense thickets after fires.

66. Ceanothus tomentosus Parry. Sierras (Amador County) southward to the San
Bernardino Mountains and San Diego.

67. Ceanothus velutinus Dougl. British Columbia to California; south in the Coast
Ranges to Marin County, and in the Sierras to Kern County; eastward to the Rocky Moun-
tains. A very important species of the conifer-forest chaparral, with its companions form-
ing forest undergrowth as well.

68. Ceanothus verrucosus Nutt. Vicinity of San Diego, southward into northern Lower
California.

CORNACE.E.

69. Garrya elliptica Dougl. Quinine bush. Coast Ranges from Oregon to the Santa
Lucia Mountains. Usually present in small number in the climax chaparral; seldom abun-
dant.

70. Garrya fremontii Torr. Inner Coast Ranges and Sierra Nevada to San Jacinto
Mountains. More closely identified with the conifer forest than with the climax chaparral.

ERICACEAE.

71. Comarostaphylis diversifolia (Parry) Greene. Southern California and northern
Lower California. Representative of a genus almost entirely Mexican and Central Amer-
ican.

72. Xylococcus bicolor Nutt. Near the coast; San Diego County and northern Lower
California. A rather important climax species within its limited range.

73. Arctostaphylos andersonii A. Gray. Manzanita. 1 Of local distribution; Oakland
Hills and Santa Cruz Mountains.

74. Arctostaphylos drupacea (Parry) n. comb. [A. pringlei var. drupacea Parry: Uva-
ursi drupacea (Parry) Abrams]. San Bernardino Mountains to northern Lower California.
Belongs to the forest region of the higher mountains.

75. Arctostaphylos glauca Lindl. From the San Francisco Bay region and Stanislaus
County to Mount San Pedro Martir in Lower California. One of the most characteristic
and conspicuous members of the climax chaparral of the southern half of the State. Gen-
erally scattered among other shrubs, often Adenostoma, and standing out conspicuously
by reason of its large size, smooth rounded form, and light gray foliage. Probably attains
a larger size than any other species of manzanita.

76. Arcloslaphrjlos liookeri Don. Near the coast, from San Francisco to San Luis
Obispo County; in the Monterey region growing with other species of restricted range.
(See No. 56.)

'The name "Manzanita" is applied to all the species indiscriminately; also to the closely
related Nos. 71 and 72.

120 THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA.

77. Arctostaphylos manzanita Parry. Mountains of the northern half of the State, from
the region of San Francisco Bay and Stanislaus County into Oregon; almost complementary
in range to A. glauca, and of equal or greater importance in the climax chaparral.

78. Arctostaphylos mariposa Dudley. Sierra foothills and to some extent in the forest
region. Of considerable importance where it occurs.

79. Arctostaphylos montana Eastwood. Central Coast Ranges, very local; on Mount
Tamalpais, where it is apparently confined to the outcrops of serpentine, which are avoided
by other species of the genus growing on the mountain.

80. Arctostaphylos myrtifolia Parry. Near lone, Amador County. Known only from
the type locality.

81. Arctostaphylos nevadensis A. Gray. Upper forest region of the Sierra Nevada; ex-
treme northern California, the mountains of Trinity County, and in the Cascades of Oregon.
A very low, spreading shrub, of considerable importance successionally in the high mountain
forest regions.

82. Arctostaphylos nummvlaria A. Gray. Outer north Coast Ranges; Santa Cruz
Mountains; Mount Tamalpais to Mendocino County. Characteristic of the Mendocino
"white plains;" not important in the chaparral.

83. Arctostaphylos parryana Lemmon. Southern Sierras and Mount Pinos to the San
Bernardino Range.

84. Arctostaphylos patvla Greene. Cascades of Oregon; eastward to the Blue Moun-
tains (northeastern Oregon) and Utah; south in California to Trinity County in the Coast
Ranges, throughout the middle altitudes of the Sierras, extending to Mount San Jacinto.
The most important species of the genus in the upper conifer-forest chaparral.

85. Arctostaphylos pumila Nutt. Endemic in the Monterey region, growing with other
local species already noted (see No. 56). A low, spreading shrub, resembling A. nevadensis
in habit; an important sand-binder on the dunes of Monterey Bay.

86. Arctostaphylos pungens H. B. K. Southern California (San Bernardino and Cuya-
maca Mountains) and Lower California to southwestern Colorado, Arizona, and Central
Mexico. Conifer-forest chaparral. An important species in the dilute chaparral of the
Santa Catalina Mountains in Arizona.

87. Arctostaphylos stanfordiana Parry. Of limited range in the inner north Coast
Ranges from Mount Diablo to Mendocino County. A rather rare species of the climax
chaparral, conspicuous by reason of its yellow-green foliage.

88. Arctostaphylos tomentosa (Pursh) Dougl. Southern British Columbia through west-
ern Washington and Oregon and the Coast Ranges and Sierras of California to Lower
California. Probably the most widely distributed member of the genus except A. uva-
ursi and possibly A. pungens. Extremely variable; several more or less well marked species
have been segregated, and the synonymy of the group is confused. An important member
of the chaparral, mainly in the climax type; frequently the only manzanita present, and
in such cases forming the bulk of the growth of the less xerophytic situations.

89. Arctostaphylos vestita Eastwood. Endemic in the Monterey region, growing with
other local species already noted (see No. 56).

90. Arctostaphylos uiscida Parry. East slopes of the inner north Coast Ranges and the
Sierra foothills; northward to southern Oregon. An important member of the chaparral
of the digger and lower yellow pine zones; forming a very beautiful gray-green cover when
in pure growth.

HYDROPHYLLACE.E.

91. Eriodiclyon californicum (H. and A.) Torr. Yerba santa. Coast Ranges and lower
Sierras. Belongs mainly with the climax chaparral; infrequent, however, in dense brush;
characteristic where climax conditions have been disturbed; frequent along roads and
trails in the chaparral. Three other species of Eriodictyon occasionally occur in the chap-
arral, but have been excluded from the category of dominants.

The Secondary Species.

In the following paragraphs the normal undergrowth of the
broad-sclerophyll forest and the climax chaparral is presented.
In addition a list of incidental species might be compiled and extended
indefinitely. Some of these are of considerable importance, one

THE BROAD-SCLEROPHYLL VEGETATION OF CALIFORNIA. 121

group particularly so. This includes species that belong properly
to the early stages of primary and secondary successions, remaining
as relicts in the climax or subclimax. They have been considered
at some length in the section dealing with development.

A complete list of the forest undergrowth is here impracticable,
since the flora is so different in various parts of the State. This is
especially true in the outer Coast Ranges, where a large element of
the herbaceous flora of the redwood forest would have to be included.
In the central Coast Ranges, away from redwood dominance, the
following shrubs are important as undergrowth:

Holodiscus discolor Rubus vitifoliua C. and S.

var. ariasfolius (Wats.) Jepson. Symphoricarpos racemosus Michx.

In the lowest stratum Aspidium rigidum var. arguium Eat. and
Micromeria chamissonis (Benth.) Greene are prominent throughout
the year, and species of Fritillaria, Calochortus, Trillium, Smilacina,
and other spring-flowering genera are conspicuous for a brief time.
A complete list from a single typical locality is given on page 38.

The herbaceous vegetation of the normal undisturbed chaparral
is exceedingly scanty, both in number of species and of individuals.
Further, few of the species that do occur can be considered as peculiar
to the chaparral. All of them could without serious inaccuracy
be consigned to the category of incidentals. However, a few have
been selected which seem to be more characteristic of the chaparral
than of any other community. All of these are perennials. Naturally
they are more abundant in the less xerophytic situations. This list
is fairly complete only for the localities with which I am most familiar.
Doubtless other species would be added with more extended obser-
vation. Several have the peculiar habit described under Zygadenus
fremontii.

List III. — Herbaceous Species of the Chaparral.

92. Gymnogramme triangularis Kaulf. Gold-back fern. Rather frequent in the more
mesophytic situations.

93. Pellaea mucronala (Eaton) Maxon [P. ornithopus Hook.]. Bird-foot fern. Char-
acteristic of especially dry situations, often where the chaparral is thin.

94. Zygadenus fremontii Torr. Frequent under the chaparral bushes in suppressed
condition, not flowering in such state; conspicuous after fire or clearing, flowering freely
and seemingly suddenly increasing in abundance.

95. Xerophyllum tenax Nutt. Turkey-beard. Decidedly infrequent, but very con-
spicuous after fires.

96. Chlorogalum pomeridianum (Ker.) Smith. Soapweed. Similar in habits to the last.

97. Brodi&a californica Jepson. Twining brodiaea. An anomalous member of the
genus in its climbing habit; straggling and weakly twining upon chaparral bushes; lower
altitudes of the Sierra Nevada.

98. Lilium washingtonianum Kellogg. Washington lily; chaparral lily. In the chapar-
ral of the Sierras and in the mountains of northern California, and northward to the Colum-
bia River.

99. Aster radulinus A. Gray. Growing in suppressed condition; flowers after clearing.

BIBLIOGRAPHY.

(1) Abrams, L. R. 1910. A phytogeographic and taxonomic study of the southern

California trees and shrubs. Bull. N. Y. Bot. Gard. 6; No. 21: 300-485.

(2) 1914. Uva-ursi. In North American Flora, vol. 29, pt. 1: 92-101. New

York Botanical Garden.

(3) and F. J. Smiley. 1915. Taxonomy and distribution of Eriodictyon.

Bot. Gaz. 60: 115-133.

(4) Adamovic, Lujo. 1909. Die Vegetations-verhaltnisse der Balkanlander. Engler

und Drude: Die Vegetation der Erde, XI.

(5) Adams, C. C. 1902. Southeastern United States as a center of geographical dis-

tribution of flora and fauna. Biol. Bull. 3, No. 3.

(6) Barber, J. H. 1898. A glimpse of the San Gabriel forest reservation. The For-

ester, 4: 240-242.

(7) Beck von Mannagetta, G. R. 1901. Die Vegetations-verhaltnisse der illyrischen

Lander. Engler und Drude: Die Vegetation der Erde, IV.

(8) Bergen, J. Y. 1903. The macchie of the Neapolitan coast region. Bot. Gaz. 35:

350-362; 416-426.

(9) Boerker, R. H. 1915. The reforestation of brush fields in northern California.

Proc. Soc. Amer. Foresters 10: 284-293; also For. Quart. 13: 15-24.

(10) Brandegee, K. 1894. Studies in Ceanothus. Proc. Calif. Acad. Sci., ser. 2, vol.

4: 173-222.

(11) Brandegee, T. S. 1889. Plants from Baja California. Proc. Calif. Acad. Sci.,

ser. 2, vol. 2: 117-216.

(12) 1891-92. The vegetation of "burns." Zoe 2: 118-122.

(13) 1893-94. The southern extension of the Calif ornian flora. Zoe 4: 199-210.

(14) Branner, J. C. 1909. Santa Cruz folio. U. S. Geol. Sur., Geol. Folio No. 163.

(15) Briggs, L. J., and H. L. Shantz. 1911. A wax seal method for determining the

lower limit of available soil moisture. Bot. Gaz. 51: 210-219.

(16) Cannon, W. A. 1911. The root habits of desert plants. Carnegie Inst. Wash.

Pub. No. 131.

(17) 1913. A note on a chaparral-forest relation at Carmel, California. Plant

World 16: 36-38.

(18) 1914. Specialization in vegetation and in environment in California.

Plant World 17: 223-237.

(19) 1914. Tree distribution in central California. Pop. Sci. Mo. 85: 417-424.

(20) 1918. The evaluation of the soil-temperature factor in root growth. Plant

World 21: 64-67.

(21) Clements, F. E. 1916. Plant succession. Carnegie Inst. Wash. Pub. No. 242.

(22) 1920. Plant Indicators: The relation of plant communities to process

and practice. Carnegie Inst. Wash. Pub. No. 290.

(23) Cooper, W. S. 1917. Redwoods, rainfall, and fog. Plant World 20: 179-189.

(24) Davy, J. B. 1902. Stock ranges of northwestern California. U. S. Dept. Agr.,

Bur. Plant Ind. Bull. 12.

(25) Diels, L. 1906. Die Pflanzenwelt von west-Australien sudlich des Wendekreises.

Engler und Drude: Die Vegetation der Erde, VII.

(26) Drude, O. 1890. Handbuch der Pflanzengeographie.

(27) Dudley, W. R. 1901. Zonal distribution of trees and shrubs in the southern

Sierra. Sierra Club Bull. 3, No. 24: 298-312.

(28) Eastwood, Alice. 1903. Notes on Garrya, with descriptions of new species and

key. Bot. Gaz. 36: 456-463.

(29) Engler, A. 1902. Die pflanzengeographische Gliederung Nordamerikas. Ab-

druck, aus dem Notizblatt des konigl. bot. Gart. Appendix IX.

(30) Foster, H. D. 1912. Interrelation between brush and tree growth on the Crater

National Forest, Oregon. Proc. Soc. Amer. Foresters, vol. 7, No. 2: 212-225.

122

BIBLIOGRAPHY. 123

(31) Fuller, G. D., and A. L. Bakke. 1918. Raunkiaer's "life forms," "leaf-size

classes," and statistical methods. Plant World 21: 25-37.

(32) Gemoll, Kurt. 1902. Anatomisch-systematische Untersuchung des Blattes aus

den Triben: Rhamneen, Colletieen und Gouanieen. Beih. zu Bot. Cent. 12:
351^24.

(33) Greene. E. L. 1893. Vegetation of the summit of Mount Hamilton. Erythea 1:

77-97.

(34) 1893. Vegetation of Mount Diablo. Erythea 1: 166-179.

(35) Grinnell, Joseph. 1908. The biota of the San Bernardino Mountains. Univ.

Calif. Pub. in Zool. 5: 1-170.

(36) Guttenberg, Hermann RiTTER von. 1907. Anatomisch-physiologische Unter-

suchungen liber das immergrune Laubblatt der Mediterranflora. Eng. Bot.
Jahrb. 38: 383-444.

(37) Haefner, H. E. 1912. Chaparral areas on the Siskiyou National Forest. Proc.

Soc. Amer. Foresters 7: 82-95.

(38) Hall, H. M. 1902. A botanical survey of San Jacinto Mountain. Univ. Calif.

Pub. in Bot. 1: 1-144.

(39) 1912. A Yosemite flora.

(40) Hansen, George. 1897. Ceanothus in landscape of the Sierra Nevada. Garden

and Forest 10: 102-103.

(41) Harshberger, J. W. 1911. Phytogeographic survey of North America. Engler

und Drude; Die Vegetation der Erde, XIII.

(42) Herzog, Theodor. 1903. Anatomisch-systematische Untersuchung des Blattes

der Rhamneen aus den Triben: Ventilagineen, Zizypheen und Rhamneen.
Beih. zu Bot. Cent. 15: 95-207.

(43) Howell, Thomas. 1903. Flora of Northwest America.

(44) Jepson, W. L. 1891. Botany of the Marysville Buttes, Sacramento Valley. Bull.

Torr. Bot. Club 18: 317-327.

(45) 1893. The mountain region of Clear Lake. Erythea 1: 10-16.

(46) 1899. Vegetation of the summit of Mount St. Helena. Erythea 7: 105-113.

(47) 1910. The silva of California. University Press, Berkeley.

(48) 1911. Flora of western middle California. 2d edition.

(49) (Date?) Regeneration in manzanita. Madrono 1: 3-11.

(50) Leiberg, J. B. 1899. San Bernardino Forest Reserve. U. S. Geol. Sur., 19th

Ann. Rep., pt. 5: 359-365.

(51) 1899. San Gabriel Forest Reserve. U. S. Geol. Sur., 19th Ann. Rep.,

pt. 5: 367-371.

(52) 1899. San Jacinto Forest Reserve. U. S. Geol. Sur., 19th Ann. Rep.,

pt. 5: 351-357.

(53) 1900. San Bernardino Forest Reserve. U. S. Geol. Sur., 20th Ann. Rep.,

pt. 5: 429-454.

(54) 1900. San Jacinto Forest Reserve. U. S. Geol. Sur., 20th Ann. Rep.,

pt. 5: 455-478.

(55) 1900. San Gabriel Forest Reserve. U. S. Geol. Sur., 20th Ann. Rep.,

pt. 5: 409-428.

(56) 1902. Forest conditions in the northern Sierra Nevada, California. U. S.

Geol. Sur., Prof. Paper 8.

(57) Livingston, B. E. 1910. Relation of soil moisture to desert vegetation. Bot.

Gaz. 50: 241-256.

(58) McAdie, A. G. 1902. Wet and dry seasons in California. U. S. Dept. Agr. Year

Book, 1902: 187-204.

(59) 1903. Climatology of California. U. S. Weather Bur. Bull. L.

(60) McKenney, R. E. B. 1901. Notes on plant distribution in southern California.

Beih. zu Bot. Cent. 10: 166-178.

(61) Merriam, C. Hart. 1893. Notes on the distribution of trees and shrubs in the

deserts and desert ranges of southern California, southern Nevada, north-
western Arizona, and southwestern Utah. U. S. Dept. Agr., Div. Biol.
Surv., N. A. Fauna No. 7: 285-343.

(62) 1899. Results of a biological survey of Mount Shasta, California. U. S.

Dept. Agr., Div. Biol. Surv., N. A. Fauna No. 16.

124 BIBLIOGRAPHY.

(63) Merriam, C. Hart. 1905. The Indian population of California. Amer. Anthrop.,

n. s., 7: 596-606.

(64) Miller, L. C. 1906. Chaparral as a watershed cover in southern California.

Proc. Soc. Amer. Foresters 1: 147-157.

(65) Moore, Barrington. 1917. The moisture withholding power of soils. Jour.

For. 15: 110-117.

(66) Munns, E. N. 1916. Results of the effect of chaparral and forest cover on meteoro-

logical conditions. Sci. 44: 759-760.
(g7) 1919. Some biological and economic aspects of chaparral. Jour, of For-
estry 17: 9-14.

(68) 1920. Chaparral cover, run-off, and erosion. Jour, of Forestry 18: 806-814.

(69) Nichols, G. E. A working basis for the ecological classification of plant com-

munities. Ecology: in press.

(70) Niedenzu, Franz. 1890. Tiber der anatomischen Bau der Laubblatter der Arbu-

toidese und Vaccinioidese in Beziehung zu ihrer systematischen Gruppierung
und geographischen Verbreitung. Eng. Bot. Jahrb. 11: 134-263.

(71) Parish, S. B. 1903. A sketch of the flora of southern California. Bot. Gaz. 36

203-222; 259-279.

(72) 1917. An enumeration of the pteridophytes and spermatophytes of the

San Bernardino Mountains, California. Plant World 20: 163-178; 208-223;
245-259.

(73) Piper, C. V., and R. K. Beattie. 1915. Flora of the Northwest Coast. State

College, Pullman, Wash.

(74) Pltjmmer, F. G. 1911. Chaparral. U. S. Dept. Agr., For. Serv. Bull. 85.

(75) Purdy, Carl. 1897. The chemise world. Garden and Forest 10: 72; 83.

(76) Purpus, C. A. T. 1897. Die Chaparral-region der siidwestlichen Sierra Nevada

von Californien. Mitt. d. deutsche dendrol. Ges. No. 6.

(77) Reed, W. G., and J. B. Kincer. 1917. Average annual precipitation map of the

United States. Mo. Wea. Rev. 45, No. 7.

(78) Reiche, K. 1907. Grundzuge der Pflanzenverbreitung in Chile. Engler und

Drude: Die Vegetation der Erde, VIII.

(79) Sargent, C. S. 1905. Manual of the trees of North America.

(80) Schimper, A. F. W. 1903. Plant geography upon a physiological basis. Transl.

by Groom and Balfour.

(81) Shreve, Forrest. 1915. The vegetation of a desert mountain range as conditioned

by climatic factors. Carnegie Inst. Wash. Pub. No. 217.

(82) Shull, C. A. 1916. Measurement of the surface forces in soils. Bot. Gaz. 62:

1-31.

(83) Smith, W. G. 1913. Raunkiaer's "life-forms" and statistical methods. Jour.

Ecol. 1: 16-26.

(84) Solereder, Hans. 1908. Systematic anatomy of the dicotyledons. Transl. by

Boodle and Frisch.

(85) Sterling, E. A. 1904. Chaparral in northern California. For. Quart. 2: 209-214.

(86) Sudworth, G. B. 1900. Stanislaus and Lake Tahoe Forest Reserves and adjacent

territory. U. S. Geol. Sur., 21st Ann. Rep., pt. 5: 499-561.

(87) 1908. Forest trees of the Pacific Slope. U. S. For. Serv.

(88) Transeau, E. N. 1905. Forest centers of eastern America. Amer. Nat. 39 : 875-889.

(89) Trelease, W. 1895-97. Rhamnaceae. In Asa Gray; Synoptical flora of North

America: 401-419.

(90) Warming, E. 1909. (Ecology of Plants. Trans, by Groom and Balfour.

(91) Wilkomm, M. 1896. Grundzuge der Pflanzenverbreitung auf der iberischen Halb-

insel. Engler und Drude: Die Vegetation der Erde I.

THE BROAD-SCLEROPHYLL VEGETATION
OF CALIFORNIA

AN ECOLOGICAL STUDY OF THE CHAPARRAL AND
ITS RELATED COMMUNITIES

BY
WILLIAM S. COOPER

Published by the Carnegie Institution op Washington
Washington. October, 1922