OCT 211^

THE FLUCTUATING CLIMATE OF
NORTH AMERICA.

By Ellswoeth Huntington.

I (From ' The Geographical Journal ' for September and October, 1912.)
i

en

THE FLUCTUATING CLIMATE OF NORTH AMERICA.
By ELLSWORTH HUNTING-TON.

(From ' Tlie Geographical Journal ' for September and October, 1912.)
Paet I. — The Ruins ot the Hohokam.
Readers who are familiar with recent discussions of changes of climate are
likely to say that the title of this article is exactly what they would
expect from an author whose name is identified with the pulsatory theory
of climatic changes. During the Pumpelly Expedition sent out by the
Carnegie Institution in 1903-4 to Transcaspia and adjacent regions the
present author came to the conclusion that in the dry regions of Central
Asia the climate of the past was distinctly moister than that of the present.
During the next two years an expedition by way of India to Chinese
Turkestan, in company with Mr. E. L. Barrett, led him to extend this
conclusion over a wider area and to believe tbat the change of climate has
not progressed regularly, but by pulsations. Still another expedition to
Palestine, Asia Minor and Greece in 1909 on behalf of Yale University
seemed to confirm the pulsatory theory, and to show that the general course
of bistory for at least three thousand years has been in close harmony with
the supposed climatic pulsations. Moreover, the observations of others,
even of men such as Beadnell, who do not believe that the climate of the
Earth has changed in recent times, seem to indicate that North Africa, on the
one hand, and Central Europe on the other, as well as Southern Europe and
large parts of Asia, have also been subject to climatic changes. Thus there
seems good ground for the conclusion that during historic times essentially
synchronous climatic pulsations have taken place in all of the vast region
of the temperate zone from China on the east, across Asia and Europe, to
the Atlantic on the west. Obviously, if such pronounced and widespread
changes have occurred in the eastern hemisphere, there is a possibility that
changes of a similar nature may have taken place in America. Accordingly
when Dr. D. T. MaeDougal, Director of the Botanical Department of the
Carnegie Institution of Washington, invited the author to co-operate with
the Desert Laboratory at Tucson, Arizona, in a climatic study of the arid
south-western portion of North America, the opportunity seemed too good
to be neglected. Three seasons, consisting of three months in the spring
of 1910, four months in 1911, and four in 1912, have now been spent in
the field. Tbe time was divided between the states of New Mexico, Arizona,
and California in the United States, and Sonora, Mexico City, Oaxaca, and
Yucatan in Mexico. Most of the methods of investigation were similar
to those whicb the writer has employed in Asia, and led to a similar result.
To these, however, were added some significant observations upon the
relation of tropical jungle and tropical forest to civilization in Yucatan,
and a series of highly conclusive measurements of trees. Both of these
lines of observation, to a remarkable degree, confirm previous conclusions.
The data concerning Yucatan and other parts of southern Mexico are to
A

265 THE FLUCTUATING CLIMATE OF NORTH AMERICA.
be published shortly in the Bulletin of the American Geographical Society.
The final results of the measurement of trees in 1912 are not yet ready for
publication. Those, together with a full account of all the results of the
three years' work, and a discussion of the bearing of climatic changes on
historical and geological theories, will soon be published in book form by
the Carnegie Institution under the title of 'The Climatic Factor in the
Evolution of Arid America.' Meanwhile the following pages contain a
resume of the chief conclusions derived from the observations of the years
1910 and 1911.
Investigations of the climate of the past may be divided into two distinct
portions dealing with the effect of climatic variations, first, upon the form
of the Earth's surface, and, second, upon the animate world, consisting of
plants, animals, and men. I shall pass over the first portion briefly,
leaving the fuller discussion for the report to be published by the
Carnegie Institution. In arid, non-glaciated regions, the effect of
climatic variations upon the Earth's surface is visible chiefly in four
ways, namely, changes in the depth and quality of the soil, the formation
of fluviatile terraces, the expansion or contraction of lakes, and the
formation or obliteration of sand dunes. When conditions become dry
there is a marked tendency toward denudation of the soil from the steeper
slopes and its accumulation in deep deposits in the valleys and on the
plains at the base of the mountains. On the return of moister conditions
the relatively greater rapidity of weathering causes rapid production
of soil, and the increase of vegetation holds this soil in place on the slopes
to a far greater degree than before. Hence the streams are not only more
steady and flow in full volume to a greater distance than formerly, but
carry a less heavy load of waste. Therefore they become agents of
erosion rather than deposition. In this way, given the proper conditions
of topography, each climatic pulsation produces a terrace. Such terraces
are found in large numbers in the arid south-western section of the United
States, for example, along the upper course of the Gila and its tributaries
in Southern Arizona. They are found almost invariably wherever the
mountains are of sufficient height and steepness. Non-climatic conditions,
such as movements of the Earth's crust, may, to be sure, produce similar
results. Nevertheless, after due allowance has been made for other causes,
the majority of the numerous fluviatile terraces which are found in such
large numbers among the mountains of the American south-west seem to
be the product of variations of climate. Their distribution, number, and
nature are almost identical with those of similar regions in Asia, which in
itself points to a common origin. The oldest may possibly go back to the
last epoch of the glacial period, but there are three or four of much later
date, and the youngest can scarcely be more than one or two thousand
years old.
Numerous old strands of enclosed salt-lakes agree with the fluviatile
terraces in indicating the prevalence of comparatively moist climatic

THE FLUCTUATING CLIMATE OF NORTH AMERICA. 266
conditions at a time long after the last glacial epoch. For instance, in the
Otero basin of Southern New Mexico certain strands are found elevated
only a score of feet above the present level of the floor of a large playa,
while others lie five to ten times as high. The lower strands show so little
weathering compared with the upper ones that they cannot be more than a
fraction as old. The upper ones probably date from the last glacial epoch,
let us say thirty thousand years ago. The younger can scarcely be more
than a tenth as old and may be still younger.
In this same Otero basin a third type of evidence of pulsatory changes
is found in a series of gypsum dunes of several ages agreeing apparently
with the strands and terraces. During arid epochs the dunes have spread
from the dry floor of the playa under the influence of strong south-west
winds. In moist times they have become fixed by vegetation, only to be
covered once more with fresh sand on the recurrence of aridity. The
agreement of these three lines of evidence — terraces, strands, and dunes —
with one another, and still more with similar phenomena in Asia, is in
itself sufficient to raise a strong presumption that changes of climate in
America have coincided with those of Asia during the last few thousand
years. Turning now to the evidence to be derived from living beings, we may
dismiss animals entirely, because their movements have not yet been
investigated in respect to the problem before us. Plants will be left for
later consideration, and we will take up man and his works. The south
western part of America is full of vestiges of pre-Columbian man. Some,
such as the cliff-dwellings and the great irrigation works and villages of
the Gila valley in Southern Arizona, are famous. A far larger number,
however, are practically unknown, and have received almost no attention
even from archaeologists. The reason is obvious. In most cases the ruins
are so insignificant that an unobservant traveller might ride miles through
what was once a region thickly studded with villages without being aware
of the fact. Walls for defensive purposes upon the mountains or picto-
graphs upon the face of the rocks are apt to attract attention, but few
people notice the far more important sites of villages scattered in profusion
over thousands of square miles, especially in Southern Arizona, New
Mexico, and the neighbouring parts of Sonora. The sites are now reduced
to barren expanses strewn with ornamented bits of pottery, flint knives and
arrow-heads, stone hammers and axes, mani and metate stones for grinding
seeds, and in some cases rectangular lines of boulders placed erect at
intervals of a foot or two and evidently outlining the walls of ancient
houses. Here and there a little mound a foot or two high shows where an
ancient dwelling was located. In almost every village an oval hollow
surrounded by a low wall covers an area 100 or 200 feet long by half as wide
— not a reservoir, as one at first supposes, but probably a ceremonial
chamber of some sort. Aside from these scanty traces nothing remains.
Yet there can be no question that these were once ancient villages.

267 THE FLUCTUATING CLIMATE OF NORTH AMERICA.
Frequently the ground is full of bits of pottery to a depth of 2 feet or
more, while the surface is so strewn with similar bits that one cannot walk
without treading on them. The houses were probably built for the most
part of branches wattled with mud. Such houses disappear quickly
when abandoned, for the wood decays, and the clay used for wattling
blows away, or else is spread over the ground in such a way as not to be
noticeable. The more well-to-do members of the community apparently
had more pretentious houses, the remains of which are probably to be found
in the larger heaps of clay and rubble which occur in most villages. Close
to the mountains, or in' regions where stone was available, other methods
of construction prevailed. There we find every type of architecture, from
houses which used stone only in a single course in the bottom of the walls,
to structures made entirely of roughly squared stone blocks. Some of these
stone structures were cliff dwellings of three stories in front of caves, while
others were isolated buildings standing in the middle of a plain and still
rising three or four stories even after the lapse of one or two thousand
years. The majority of the villages must have been inhabited for a long
time. Even where the houses have entirely disappeared, the amount of
broken pottery covering the ground indicates that a busy population lived
here for centuries. In certain cases, to be sure, the quantity of pottery is
so small as to indicate only a brief occupancy. In the larger villages,
however, the amount is literally scores of times as great as in certain
modern villages which I examined. For example, Cababi and Juwak, near
the southern border of Arizona, are villages which have been occupied by
the Papago Indians for at least fifty years. So slow are these people to
adopt modern methods that they still use pottery to practically the same
extent as before the coming of the White Man. The amount of broken
pottery in such villages is insignificant, while that in the ruins is as great
as in many Asiatic ruins which are well known to have been occupied
hundreds of years. I emphasize this.point because American archseologists
and ethnologists have laboured under a peculiar impression which amounts
almost to an hallucination. Being convinced that no change of climate
has occurred, they have been forced to the peculiar theory that the ancient
people of America, the " Hohokam " or " Perished ones," as the modern
Pimas called them, were of a different nature from the rest of mankind.
It has been supposed that these ancient Hohokam were extraordinarily
mobile and extraordinarily industrious. For instance, in the thirteenth
annual report of the " Bureau of Ethnology " (p. 259), Mindeleff, one of the
best authorities, say that " a band of five hundred village-building Indians
might leave the ruins of fifty villages in the course of a single century."
He assumes a degree of mobility absolutely unparalleled among any modern
agricultural people, or among any of whom we have historic records. His
assumption also carries with it the corollary that the Hohokam must have
spent most of their time building houses or in making pottery with which

THE FLUCTUATING CLIMATE OF NORTH AMERICA. 268
to strew the ground and give an appearance of age to their villages.
Hunting tribes are of course mobile in the highest degree, but the people
with whom we are dealing were strictly agricultural, as is universally
agreed. The ruins of their villages are invariably located close to agri
cultural land, or at least to land which would be available for agriculture
if there were water enough ; their number, even according to those who
hold the migratory theory, was too great to allow of their obtaining a
living by hunting ; they had no domestic animals on which to rely ; and
finally, traces of corn and beans, the two staple products, are found in
almost every ruin. Accordingly, in the following discussion we shall
follow the archaeologists in assuming that the ancient Hohokam were an
agricultural people. We shall depart from them, however, in assuming
that, in the absence of any evidence to the contrary, the Hohokam were
like the rest of mankind, and their ruins are to be interpreted by the same
criteria as those universally employed in the study of the archaeology of
other parts of the world.
With these assumptions in mind, we are prepared to investigate the
relation of the ancient population to rainfall. Let us first concentrate
our attention upon a single region, the Santa Cruz valley of Southern
Arizona. I select this valley, not because it is particularly remarkable,
but because it happens to contain Tucson, the site of the Desert Laboratory.
This town is the largest in the two states of New Mexico and Arizona,
although it has only sixteen thousand people, including all its suburbs.
The reason for the scantiness of population is found in the climate.
The average rainfall at Tucson amounts to about 12 inches. This is
distributed between two rainy seasons ; one of them comes in the winter
from November to March, and has an average of about 5 inches of rain,
while the other, with 7 inches of rain, begins at the end of June and lasts
until early in September. The months of April, May and June, or the
foresummer, as MacDougal has called them, are practically rainless and
very hot, and the same is true of the interval from the end of the summer
rains to the beginning of those of winter. Nothing can be raised without
irrigation of some sort. Since the coming of the White Man, winter crops,
such as barley, alfalfa, and the like, have become important. The Indians,
however, cultivated practically nothing except corn and beans, which they
irrigated by means of the summer floods.
The Santa Cruz valley has a length of at least 200 miles, but most of
it is well-nigh a desert, and can be utilized only for cattle -raising.
According to Prof. R. H. Forbes, Director of the Arizona Experiment
Station, the entire drainage area of the Santa Cruz river contains only
about 6000 acres of land under cultivation. Part of the 6000 acres is
under full irrigation and produces four or five crops of alfalfa per year,
while a considerable portion is only under flood irrigation and produces
but one crop each year, when the heavy rains of July and August redeem
the desert for a brief space. Under the best system of irrigation available

269 THE FLUCTUATING CLIMATE OF NORTH AMERICA.
in modern times Prof. Forbes estimates that for every two acres brought
under full cultivation one person is added to the population of Arizona.
In other words, if the Santa Cruz valley wore cut off from all the rest of
the world and left to its own resources without railways, mines, health-
seekers or other extraneous sources of wealth, the population would be
limited to the three thousand who could be supported on the 6000 acres of
irrigated or partly irrigated land. To this number nothing could be added
by dry farming without irrigation ; for Prof. Forbes expressly states that
at the present time, in spite of various attempts, no such thing as genuine
dry farming exists in the state of Arizona. Promising experiments give
hope of some success in the future, but they involve repeated and expensive
ploughing of the soil after each short period of rain, and this must be kept
up for two years before any orop can be harvested. It may safely be
assumed that the ancient Hohokam, with no iron tools, no beasts of burden,
and no great knowledge of. science, could scarcely cultivate as much land
as the modern American. Moreover, as they had no winter crops, they
could scarcely have raised as much food per acre as is now possible, even
had they not been otherwise handicapped. For the sake of argument,
however, let us suppose that with the aid of wild fruits and seeds in bad
years, and with their lower standards of living, the Hohokam, if they were
here to-day under the present conditions of climate, might support them
selves to a maximum number of four or five thousand.
Granting that four or five thousand Hohokam might possibly find a
living in the Santa Cruz valley under present climatic conditions, let us
see where they would be located. Inasmuch as the ancient people were
agricultural, they must have lived where both land and water were avail
able. At present about 1500 of the 6000 arable acres are at the Indian
reservation and old Spanish mission of San Xavier, 9 miles up the Santa
Cruz to the south of Tucson. Six or seven hundred Indians now live there,
cultivating the land, raising cattle, and going out to the neighbouring city
to work. In the days of the Hohokam a somewhat dense population lived
at San Xavier, as is proved by various ruins, including a fort on a hilltop.
Around Tucson itself the modern houses and streets make it impossible to
determine exactly how large an area was ocoupied by the Hohokam. In
all the outskirts of the town, however, pottery and other evidences of
early man are abundant, and there is a fort on Tumamoc hill where the
Desert Laboratory is now located. Evidently many Hohokam lived near
Tucson and cultivated the 2000 acres which can there be irrigated.
A third large area of modern cultivation is found along the Rillito, a
stream from the south-east which joins the northward flowing Santa
Cruz, 8 miles north of Tucson. Here nearly 2000 acres are now in use.
In the past the Hohokam evidently made use of the same land, for
traces of villages are found near Agua Caliente, Tanke Verde, Fort
Lowell, and elsewhere. The three areas of San Xavier, Tucson, and tbe
Rillito valley include about 5500 out of the 6000 acres available for

DUNES OF PUBE GYPSUM NEAB THE DESICCATED LAKE OF OTEEO, IN SOUTHEEN
NEW MEXICO.

THE SITE OF A WATEELESS BUIN IN THE SANTA CEUZ BASIN AT RINCON. THE BOWS
OF STONES IN THE FOBEGBOUND AND EXTENDING TO THE LEFT FEOM THE
MEXICAN ABE PABTS OF THE FOUNDATIONS OF ANCIENT HOUSES.

Xf^C/-

."%

M

TEERACES APPARENTLY FOR CULTIVATION ON A DRY HILLSIDE AT TEINCHEBAS. THE
PROMINENT RECTANGLE IN THE CENTRE APPEARS TO HAVE BEEN THE SIDE OF
A RELIGIOUS STRUCTURE.

TYPICAL FOREST OF SUHUARO OR GIANT CACTUS IN THE SANTA CRUZ VALLEY.

THE FLUCTUATING CLIMATE OF NORTH AMERICA, 270
cultivation ; while the remaining 500 are scattered here and there in
insignificant patches.
I have described the present distribution of population in the Santa
Cruz valley, and its relation to ruins, in order to bring out the fact that the
prehistoric inhabitants utilized every site which their modern successors
can utilize. They did more than this, however, for in the now almost
uninhabited and waterless 50 miles below the main irrigated areas of
Tucson they occupied at least seven distinct villages, and others will
probably be discovered. The first village, at Jaynes station on the
Southern Pacific railroad, 7 miles below Tucson, must have been a genuine
town. Broken bits of pottery, old grinding stones, pestles, stone hammers,
flint arrowheads, and the low mounds of old houses extend for over a mile
along a slight ridge of gravel between two areas of low-lying fertile land
easy to irrigate and cultivate if the river contained water. Now, however,
the river does not flow so far except in floods. It might possibly reach
this point permanently if no water were taken out upstream, but this is
not certain, and we have already seen that in former times there were
numerous villages farther upstream which must have used up a large amount
of water. The abundant traces of human occupation in the Jaynes village
appear to indicate that the houses were close togother, and were occupied
hundreds of years. This one village can scarcely have had less than one
or two thousand inhabitants. Adjoining it on the east we traced what
seemed to be the line of a canal for more than a mile on the gravelly
ground between the bottom lands of the Santa Cruz and the Rillito. Here '
we found not only numerous remnants of houses, but several of the hollows
described above which seem to be ceremonial chambers, and which by
their size and number point to a somewhat large population.
Two miles downstream from the Jaynes ruins, at the Nine-mile Water-
hole, where the last permanent spring is now found in the dry river-bed,
another village was located, not so large as its neighbour, but nearly half
a mile long. On the opposite or north side of the Rillito we found
traces not exactly of a village, but of a series of houses scattered along the
edge of the arable land at intervals of a few hundred feet. If all these
ruined sites were occupied at one time, which was probably the case, as
we shall see later, the population of this one small region from 7 to 10
miles below Tucson can scarcely have been less than two or three thousand.
That the inhabitants cultivated the low land on all sides of them can
hardly be doubted ; for the houses are located just on the edges of the
good land, but without encroaching upon it except where the arable tracts
are so extensive that an elevated, gravelly site cannot be found within a
reasonable distance. The villages in the situations thus far mentioned
might at present succeed in getting drinking-water without much difficulty.
They could scarcely get a living from agriculture, however — and certainly
not from any other source — -unless the upper villages were unoccupied and
all the water normally used by them were brought many miles downstream

271 THE FLUCTUATING CLIMATE OF NORTH AMERICA.
in canals. This is scarcely possible, for the universal tendency in all
irrigated regions is to use the water as far upstream as can be done
profitably under the existing state of skill iu the science of irrigation.
In the good years, to be sure, the people of these villages might raise
crops, but in the dry years they would starve.
Farther down the Santa Cruz river, by which I mean the dry bed wherein
a little water occasionally flows, a small ruin lies at the mouth of the Canada
del Oro, or Little Canyon of Gold. A little farther downstream, at the
so-called Point of the Mountains, 17 miles north-west of Tucson, there
is another large ruin known as Charco Yuma. In the spring of 1910 we
found that the nearest source whence people like the Hohokam, who
could not dig deep wells, could have obtained water was 8 miles up
stream at the Nine-mile Water-hole. There the amount was sufficient for
drinking purposes, but not for irrigation. Ranchers engaged in raising
cattle informed us that no water whatever had come down the river during
the preceding winter, although during the summer of 1909, when the
rainfall for the hot season amounted to exactly the average amount of 7
inches, floods came down after fifteen or twenty showers. In some cases
the flow of water lasted only two hours ; in the height of the rainy season,
however, that is at tbe end of July and beginning of August, a brook of more
or less size flowed steadily for two weeks. The average duration of tho
floods was said to be about thirty-six hours. From this we infer that during
a summer of average rainfall, surface water flows as far as the old village of
¦ Charco Yuma for twenty-five or thirty days during July and August. This
conclusion is confirmed by the statements of Socoro Ruelas, a Mexican cattle
rancher, who in boyhood and early manhood lived at the old stage station
located in the midst of the ruins. In winter, according to his statement,
water rarely reaches the place, and even the heavy showers of summer some
times fail to send it so far. When he was a boy in the late seventies or early
eighties the spring at the Nine-mile Water-hole increased so much as to send
out a stream that was used for irrigation for a year or two. But at other
times it completely dried up, so that there was no surface water within about
12 miles of Charco Yuma. From the spring of 1885 to August, 1887,
according to the dates given by the Mexican, no water whatever, either in
summer or winter, came down as far as the ruins, and people like the
Hohokam who depend upon summer floods would have had no crop what
ever in 1885 and 1886, and only a poor one when the late rains of 1887
arrived. A Mexican family could live there because, having dug a well in
1884, they had found water at a depth of 28 feet. During the following
dry years the level fell several feet lower, but never absolutely failed. As
the Mexicans depended on cattle and not on agriculture, the absence of
surface water did not absolutely prevent them from making a living,
although the scarcity of forage caused them much distress.
In spite of what has just been said, there is a little cultivation on the
opposite side of the Santa Cruz valley, a mile north of the ruins of Charco

THE FLUCTUATING CLIMATE OF NORTH AMERICA. 272
Yuma. Here, near the villageless station of Rillito on the Southern Pacific
railroad, a narrow strip of land containing some 300 acres and extending
along the railroad for about 3 miles is cultivated by the stationmaster and
a neighbouring cattle rancher. " Talk about dry farming,'' said the station-
master, " it's the easiest sort of thing. Five inches of rain a year is all
we need here, and we get on an average twelve. Just look at my fields.
They're not so good as usual, but they show what can be done in a bad year
like this. It's all in the way you plough and harrow and roll." A little
investigation, however, sufficed to show that the 300 acres are supplied
with very effective irrigation, not artificial but natural. At this point moun
tainous spurs cause the broad waste-filled basin of the Santa Cruz to contract
to a width of only a little over a mile. The rock floor of the basin lies near
the surface and acts as a concealed dam to raise the level of the ground water.
Also, the floods, when they come so far, spread out here and accumulate
in pools. The whole tract is therefore uncommonly moist. At the time of
my visit the barley in the low places where water sometimes stands was
much more flourishing than elsewhere ; while in wet years, according to the
stationmaster, the standing pools prevent planting. The crop was poor when
I saw it because, although the rainfall of the winter of 1909-10 was only a
little less than the average, it was badly distributed, most of it falling early
in the winter. The grain planted in September and October and even in
early November grew fairly well, while that planted later failed to head.
Even in the best part of the 300 acres, however, that is in the most moist
depressions, the hay crop, for which the barley is planted, was expected to
amount to only about 15 tons as compared with 95 in the preceding year.
What sort of a crop would be harvested in a really bad year no one
knows, for cultivation has only been carried on a few seasons. At least
a quarter and probably a third of the winters during the last forty years,
since records of precipitation began to be kept in the region, have been
even more unpropitious than 1909-10. If a poorly distributed but not
especially small rainfall in that year could cause the diminution of the
crops to the extent of five-sixths, it requires no demonstration to show
that the fields must have been almost useless in the nine years since 1887
when the winter rainfall has been less than 2-5 inches. Outside of the
300 acres now in use many attempts at cultivation have been made near
Charco Yuma, but without success. In years like 1904-5, to be sure, with
nearly 15 inches of winter rain between October and April and 6 in the
summer, from July to September, or like 1906-7 with nearly 8 in winter and
11 in summer, fine crops can be raised in many places, but this is the ex
ception, not the rule.
To sum up the conditions at Charco Yuma, it appears that no permanent
supply of water is available, without the digging of wells at least 25 feet
deep, a task impossible for a primitive people without iron tools. The nearest
permanent supply of water is 8 miles away, and even that occasionally
fails. A period of two years and more may pass without a single temporary
A 2

273 THE FLUCTUATING CLIMATE OF NORTH AMERICA.
flow of water. The total amount of land capable of cultivation amounts to
about 300 acres sufficient to support 150 or 200 people in ordinary years,
but the yield from this land has fallen off as much as 85 per cent, in recent
dry years, and must fall off more in the frequent seasons which are worse
than the one in question. Winter crops such as have just been mentioned
however, had no importance for the Hohokam since they had uo wheat,
barley, oats or other grains of the Old World, but depended almost solely
upon maize and beans which require summer rain. According to Ruelas,
the Mexican, no flood waters reached Charco Yuma in 1885, when the summer
rainfall amounted to only 3*07 inches, the minimum on record, nor in
1886 when it amounted to 4-27 inches. If no water reached the place with
a rainfall of 427 inches we may safely infer that at least 5 inches would
be required in order to raise any crop whatever. During the 42 years for
which records are available, 15 summers, or more than one-third, have had
a rainfall of less than 5 inches. We seem compelled to conclude that at
present the total amount of land which could be cultivated under the
Hohokam methods amounts to only 300 acres, which would yield no
appreciable crop at least one year out of three, and poor crops about half
of the rest of the time.
In spite of these untoward circumstances the Hohokam lived here in
considerable numbers. On the left bank of the dry river-bed, between
the channel and the base of the jutting point of the Tucson mountains,
Mr Herbert Brown, editor of the Tucson Star, showed us the remains of
a large village. For nearly 2 miles we found pottery and other artifacts
scattered along the base of the mountains, not thick as a rule, but at
frequent intervals as if houses had been located here and there along the edge
of the cultivated land just as we have seen to be the case farther upstream,
or as the modern houses of the Papago Indians are located at San Xavier.
In the centre of the village the pottery is thicker. There we found a
great boulder of andesitic lava almost buried in alluvium,, It was studded
with 24 round holes about 10 inches deep and 3 or 4 in diameter, while
a similar block not far away contained 7 holes of the same sort. Long ago
the Hohokam women must have gathered here with their stone pestles,
and gossiped as they sat on the great rooks and pounded corn, beans, or
other seeds to make flour for the daily bread of their husbands and sons.
Not far away an elliptical enclosure of the kind which we have sup
posed to be a temple or place for religious ceremonials, has a length of
210 feet and a width of 90, dimensions sufficient to indicate a village of
considerable size. Back of the temple and the great grinding stone, if
these are the proper terms, the whole eastern and northern face of the steep
rocky hills is covered with low defensive walls, enclosing spaces 10 to 30
feet wide, where families appear to have taken refuge in times of danger.
At first one is inclined to say that there must be a thousand of them.
We did not actually count, but a rough estimate shows that they surely
number several hundred, which gives some idea of the probable size of

THE FLUCTUATING CLIMATE OF NORTH AMERICA. 274
the village down below. In addition to the ruins already described there
was still another large village in this vicinity on the farther side of the
Point of the Mountains, that is half or three-quarters of a mile downstream.
Here for nearly a mile and a half along the gravel terrace above the
alluvial plain of the Santa Cruz, and overlooking at a distance the 300 acres
now under cultivation, pottery and the usual accompanying artifacts are
thickly scattered. The central part of the village occupies an area of
about 200 acres, while the surrounding part where population was less
dense covers a slightly larger area. In the centre of the village, pottery
is very thickly strewn, and the upper layers of earth are full of it to a
depth of 2 feet indicating very long occupation. This lower village
was larger than the one on the other side of the mountains; and, to
judge from the degree of ruin of its ceremonial chambers, it appears
to have been occupied longer and abandoned earlier. Moreover, there
are no defensive walls, so far as we could discover on the side of the
hills facing it, although there are plenty on the other side. All these
things suggest that the original village was down stream from the Point
of the Mountains, and existed during a time of peace, prosperity and
dense population, while the other village was occupied later, perhaps at
a time when the water-supply was becoming less so that it was necessary
to move upstream above the rock-dam of the mountains. At the same
time more war-like conditions began to prevail, possibly because of the
distress due to lack of rain ; and accordingly the people of the village were
obliged to build places of defence upon the mountain-side.
Below Charco Yuma agricultural conditions become less and less
favourable. Nevertheless ruins of villages are found at Nelson's Desert
Ranch 26 miles from Tucson, at Picacho 16 miles farther down the Santa
Cruz, and in several localities south of Toltec Station which is about 55
miles north-west of Tucson. These last sites lie in a region where a
recent attempt at irrigation has been made on a large scale. A dam of
earth was thrown up to collect flood-water, and canals were constructed to
distribute it over a plain of the richest soil. My first knowledge of the
project came from the sight of a cloud of dust which I saw from a moun
tain-top many miles away. " That," said my companion, " is the dust of
the 4000 acres that the Santa Cruz Reservoir Company ploughed up and
planted with barley last fall. Not a grain of their barley has come up in
the whole 4000 acres, and yet they are trying to sell farms." Later I rode
through the dust of the great ploughed field — dust ankle-deep — and saw
with my own eyes that there was not a vestige of any growing thing.
During the winter no water came down to fill the reservoir ; the next
summer brought no better success, and the project was abandoned. The
entire group of ancient villages in this region of unsuccessful irrigation
below Charco Yuma must have contained as many people as the groups at
either Jaynes or Charco Yuma. In every village of this most remote
group the problem not only of raising crops, but of obtaining drinking-

275 THE FLUCTUATING CLLMATE OF NORTH AMERICA.
water, would utterly prevent primitive people from living there under
present conditions. Wells a hundred or more feet deep, such as support
the few far scattered cattle ranches of modern times, were of course out of
the question. Reservoirs were doubtless the common resource, but they
were apparently constructed merely of earth without plaster, and at best
they must have been shallow. If for ten consecutive months no floods came
down the river to this point, as happened in 1909-10, all the water must
have disappeared by seepage or evaporation, to say nothing of daily use, long
before the supply was replenished ; while periods like 1884-6, when there
were no floods for at least thirty months, would inevitably cause the com
plete abandonment of such a village and probably the death of most of its
inhabitants, by reason, not only of hunger and famine, but also of the wars
and dissensions, which would inevitably arise when all the country was in
the throes of terrible drought.
To sum up the conditions in the Santa Cruz valley, it appears that
at present not more than four or five thousand people could find susten
ance without modern railroads and other means of outside assistance. The
part of the valley which is now capable of cultivation contains ruins which
indicate that all the available land was utilized in the past. Below the point
where irrigation is now possible there are three large groups of ruins, and
the three together must have had as many people as the higher regions where
there is still water. In other words, it seems as if the Santa Cruz valley
once had at least twice as many people as it could at present support, and
half of these lived where the white man cannot now get a living from
agriculture. Before leaving the Santa Cruz drainage area we must describe two
sites located at the headwaters of tributaries and affording phenomena
different from those thus far discussed. The first is at Gibbon's ranch,
east of Tucson, at the southern base of the Santa Catalina mountains.
This is one of the few places where the water-supply depends upon a
spring rather than a stream. In 1910 we found the site unoccupied,
although a decaying adobe house stands beside a small reservoir supplied
by two or three trickling little springs. The total amount of water in
March, 1910, would scarcely have sufficed to irrigate three or four acres
an amount so small that the owner of the ranch did not find it worth while
to practise agriculture at all, but turned his attention to cattle raising.
Since his death or removal, no one has lived there. Some day, perhaps, a
thrifty Chinese peasant will establish there a market garden. He will
certainly be most skilful if he can make the water suffice for the support
of more than two or three families. Yet once there was a respectable
village here. East of the dry " wash " or sandy flood channel which
occasionally carries water for an hour or two while rain is falling, my
guide, Mr. Bovee, and I counted the stone foundations of twenty-eight
houses in an area of 8 acres, while we found pottery quite thickly strewn
over 40 acres. We doubtless missed some of the houses, either because

CLIFF DWELLINGS IN THE CANYON DE LOS FRINOLES.

RUINS OF TUIYONI IN THE CANYON DE LOS FBINOLES.

CLIFF DWELLINGS AT THE LABGE RUINS OF PUYE.

PREHISTORIC CARVINGS AT PUYE.

THE FLUCTUATING CLIMATE OF NORTH AMERICA. 276
their foundations have been concealed by the washing of gravel in floods,
or because their appearance is insignificant. Hence we seem to be well
within the limit if we say that the population of the village in its prime
was probably as much as 120, whereas now the available water supply is
sufficient for little more than one- tenth as many. I am well aware that
the common objection to this conclusion is that the ancient Americans did
not occupy all the houses of a village at once ; that people built houses,
and then if some one died they built others ; and thus there might be
many times as large a number of houses as there were families. This
view, however, is absolutely unproved. Its basis is, first, the fact that
certain modern tribes of Indians, few in number, and living largely in
huts of a temporary and easily constructed character, have the habit here
described ; and, second, the unfounded assumption that the population of
the south-west in ancient, prehistoric times cannot possibly have been so
large as the ruins would seem to indicate because the country cannot now
support so many people. In reply to this, it may be said, first, that we
have absolutely no knowledge of the customs of the Hohokam, apart from
what we learn from their ruins, nor do we even know that they were
related to the modern Indians any more closely than we are supposed to
be to our fellow Indo-Europeans, the Persians. In the second place, the
habit of abandoning dwellings after some one has died in them is not at
all common at the present time, and it rarely, or possibly never prevails
among a purely sedentary, agricultural people such as the Hohokam appear
to have been. It is practically limited to tribes who wander from place
to place and whose habitations are consequently of the nature of booths or
mere temporary shelters, easily removed and easily renewed. In the
third place, the size of the central structure of each ruined village, the
supposed ceremonial chamber, is sufficient to indicate a considerable
number of people. The stone foundations of the structure at Gibbon's
Ranch, for example, have a length of 105 feet, and it hardly seems
probable that a mere handful of Hohokam, ten to twenty in number
including women and children, would have built so large a building.
Finally, it is in the highest degree unscientific to assume that conditions
of climate have always been as they are to-day. It may be that those who
hold this view are right, but they are certainly not justified in making the
assumption when the matter has never yet been submitted tD any rigorous
mathematical investigation such as is here attempted.
After studying the ruins of the main Santa Cruz valley I came to the
conclusion that a mere examination of the map was sufficient to indicate
where ruins would be found. Accordingly I decided upon the head of the
Rincon valley, about 22 miles south-east of Tucson, as a test case. The
expected ruins were found in the shape of several small villages, the chief
of which contains the foundations of at least eighteen houses. Probably
there were once many other houses in the valley, for abundant pottery is
found in several places where no foundations are visible. The present

277 THE FLUCTUATING CLIMATE OF NORTH AMERICA.
population consists of two Mexican and two American families, one of the
latter being the forest ranger ; but the arable land is said to amount to
150 or 200 acres, which would permit of a population of twenty or twenty-
five families. The significant fact in this case is not the discrepancy
between the present and past population. The number of inhabitants
in the past was not limited by the amount of water available in the stream,
but by the amount of level land in the narrow bottom of the valley.
Apparently the Hohokam wanted more land than they could readily obtain.
About a mile and a half east of the forest ranger's house and some 3 miles
east of the prominent hill called Sentinel Butte, a grassy slope drops
towards the north-west at the base of Rincon peak, 8465 feet high. Tho
slope has a fall of about 10 degrees, and an altitude of from 3300 to 3500
feet above sea-level. On the most favourable part of the slope, for a
distance of about half a mile parallel to the upper Rincon, and for a width
of half or two-thirds as much, one finds unmistakable terraces built
apparently for purposes of agriculture. In general they are from 20 to
70 feet long and 2 or 3 feet high. The commonest location is at right
angles to the minor drainage lines, each little swale being broken into
terraces with a width of from 20 to 30 feet. In some cases the spaces
between the swales are also terraced, the terrace walls in all cases being
composed of pebbles and cobblestones. I searched carefully for pottery,
but succeeded in finding only one or two coarse bits, quite in contrast to
the abundant potsherds which occur not only among the foundations lower
down the valley, but along the borders of the alluvial plain. The only
other works of man among the terraces are some small stone circles like the
beds where the modern Indians cook the yucca to make the drink known as
mescal, and the low round structure of boulders. The whole hillside is
almost exactly like hundreds in Palestine, Syria, and Asia Minor, or like
others in Mexico and South America. Even the little round structure, about
7 feet in diameter, with its small doorway, resembles the watchmen's shelters
in the terraced fields of Syria. There can scarcely be any doubt that the
terraces were designed for agriculture. Apparently they were not intended
for irrigation, for they are not properly arranged for this, nor does there
appear to be any available source of water. They must have been intended
for dry farming. Three or four miles down the valley, near the main
village, a few similar terraces occur on a gravel slope east of the ruins
and far above any available supply of water, so that here too dry farming
appears to have been attempted.
Prof. Forbes, of the Arizona Experiment Station, as has already been
stated, says that dry farming is not now practicable in Arizona except by
most careful and expensive methods of ploughing and harrowing continued
for two years in order to get a single crop. The terraced slope in the
upper Rincon valley, because of its proximity to the mountains, un
doubtedly receives more rain than do many parts of the country. Over
on the east side of the neighbouring Santa Rita mountains at an elevation

THE FLUCTUATING CLIMATE OF NORTH AMERICA. 278
considerably greater than that of our terraces, potatoes have been reported
as grown without irrigation, but inquiry shows that they are watered
naturally by springs seeping out above them. In the same region, at the
mouth of Gardener's canyon at an elevation of 5000 feet, four or five
settlers took up land and attempted real dry farming in 1909. The eleva
tion is sufficient to ensure moderately cool weather, and hence less evapora
tion than in the parching plains. The rainfall during the summer season
of 1909, as measured at the Empire Ranch not far away, amounted to
9"39 inches as against an average of 7-93 for the preceding fifteen years.
Nevertheless, the corn of the settlers failed utterly ; and the beans, the
most reliable of all crops, gave so scanty a return that the farmers were
completely discouraged. In the absence of records we cannot say cate
gorically that crops might not be raised on the terraces at Rincon. We
can merely say that nothing of the kind has succeeded in this part of
Arizona hitherto, and that in April, 1910, we found the terraces with no
more sign of fresh vegetation than was apparent on the surrounding dry
plains. The hypothesis may be advanced that the Hohokam prepared the
terraces for the cultivation of some special crop, such as wild tobacco or
some plant now unknown, for which dry conditions are especially favourable.
Wild tobacco does not now grow here, however, and we know of no other
plant which could have fitted the requirements. It is scarcely probable
that a plant of sufficient importance to demand so large a space and so
great an expenditure of effort would have escaped from cultivation and
from knowledge after having once become a part of man's equipment.
Apart from the immediate question of the possible climatic significance
of the terraces, they are important in another aspect. Mankind rarely
labours except under the compulsion of some strong force such as hunger,
fear, or ambition. The ancient Hohokam would scarcely have gone to the
labour of making the terraces without some good motive. The obvious
agricultural character of the terraces precludes the possibility of any
religious significance, as does the fact that elsewhere such terraces are
found closely associated with religious structures from which they are
clearly differentiated. Defence apparently had nothing to do with the
matter, for there seems to be no fortress near at hand, the terraces are in a
decidedly undefensible location, and they are not protected by walls. The
only adequate explanation would seem to be the need of more abundant
areas of cultivation. Probably the Hohokam of the Rincon valley found
that the limited amount of land in the bottom of their narrow valley was
not sufficent for their needs. Therefore, having somehow learned the art
of making terraces as practised in other parts of the arid south-west or in
Mexico, they built a considerable number, partly close to their main
village, but chiefly on a slope of especially favourable location. This in itself
may seem of small importance, but it is significant as indicating that
probably the population was decidedly dense. Had there been abundant

279 THE FLUCTUATING CLIMATE OF NORTH AMERICA.
unused irrigable land either in the Rincon valley or in neighbouring regions,
the Hohokam would scarcely have gone to the labour of building terraces.
Hence we infer that the population of the country as a whole was as dense
as is indicated by the abundant ruins.
The preceding pages have been devoted to the description of the
phenomena ofa single valley in southern Arizona. That valley was not.
chosen because of the strength of its evidences of climatic changes, but
merely because it happened to be the first which I investigated and the one
where I spent most time. Half a dozen others might have been chosen
equally well. The same type of phenomena are displayed with equal clear
ness in the Altar valley of north-western Mexico, in the Chaco valley 500
miles to the north-east in the north-western corner of the states of New
Mexico, and in unnumbered valleys between the two, and in the regions
round about. Wherever I investigated the matter I found the same type
of evidence. Further details would overcrowd this article and I must dis
miss this phase of the subject with one more brief example. The Jarilla
mountains are an insignificant group of detached hills lying in the centre
of southern New Mexico, 50 miles north-north-east of El Paso and 300
miles east of Tucson. Because of the presence of copper and other metals
a small mining industry grew up here a few years ago, but its boom is over
and only a few hopeful prospectors still remain. As no water could be
obtained among the mountains, the railroad had previously been forced to
construct a pipe-line to some higher mountains 20 miles to the east. Except
during the rainy seasons when a little water is stored in cisterns, the water
piped from the east has been the sole supply, not only for the railroad itself,
but for the two little mining towns so long as they had any inhabitants.
So far as I could learn, only two persons, both of whom combine a little
cattle-raising with the enticing but unremunerative occupation of prospect
ing, depend on any other water supply. These two men live far off from the
rest of mankind, and get their water from deep wells, the digging of which
would be utterly beyond the capacity of primitive people without iron tools.
The only surface water is in Water canyon, but the name is a misnomer.
The settler who lives there was most scornful when I asked about the
"spring." He said there was no spring, only a damp spot in years of unusually
good rainfall. My visit to the place confirmed his statement. I saw nothing
but a waterless valley, slightly damp because rain had fallen the day before ;
but no one would ever suspect the presence of a spring except from the
traces of an old path and of an Indian settlement. In spite of the present
absolutely uninhabitable character of the Jarilla mountains for any people
not able to dig wells, and in spite of the fact that no crops whatever can be
raised there without constant irrigation, I saw the remains of three distinct
villages of the kind already described. The pottery and other relics of
man are not so thick as in the large villages of the Santa Cruz, but they are
so abundant that the ground is thickly strewn with them. Not one of the
villages is less than half an hour's walk from the dry spring, and two of

THE FLUCTUATING CLIMATE OF NORTH AMERICA. 280
them are 4 or 5 miles away. All are obviously located close to land
which could be cultivated by flood irrigation if there were enough water
and if the inhabitants could have a permanent supply to drink. The case is
only one of hundreds scattered all over New Mexico and Arizona. It
seems inexplicable on the theory of climatic uniformity, but perfectly
reasonable if the climate of the past was different from that of the present.

a 3

392 THE FLUCTUATING CLIMATE OF NORTH AMERICA.

Paet II. — The Succession of Civilization.
In the preceding part of this article we have dealt solely with the question
of whether the climate of the past was different from that of the present.
Having concluded that there is strong evidence that this was the case,
we shall now attempt to ascertain whether the change from the past to
the present took place gradually or in pulsatory fashion, and whether its
various phases synchronize with similar phases in the Old World.
Alluvial terraces are the first matter to be dealt with in this connection.
We have already seen that their formation appears to indicate oscillations
between dry conditions when soil and rock waste were washed down from
the mountain sides into the valleys, and moist conditions when the deposits
formed in the valley bottoms were dissected because the load of the streams
had been diminished. In various places pottery or other traces of human
occupation have been found in the terraces at a considerable depth. For
instance, at Rosemont, south-east of Tucson, Mr. E. T. Vail, when digging
a well for a windmill, found potsherds at a depth of 30 feet. None what
ever are now found upon the surface in that particular vicinity. The
depth at which they are buried is so great that neither animals nor
primitive man could have carried them down from the present surface.
The only possible explanation seems to be that the level of the valley
bottom was 30 feet lower in the days of the Hohokam than at present,
and has since been raised by alluvial deposits. In other places similar
phenomena are found. For example, on one of the northern tributaries of
the Gila, Mindeleff found the remains of an old canal which had been
buried under 10 or 12 feet of silt, after which it was again exposed by
the cutting of a new stream- channel. At present the stream has cut to a
level considerably lower than that of the old surface upon which the
Hohokam evidently carried on cultivation. In both northern and southern
New Mexico I found similar instances where the habits of the streams
have changed once or twice since the days of the primitive inhabitants.
These cases are analogous to that of Olympia which was discussed in this
Journal not long ago. At the time of the occupation of the old villages,
and of the construction of the canals, the streams probably flowed at a
level slightly lower than that where the pottery was found, for the ancient
villages were presumably free from the danger of floods. Then there came
a time, apparently characterized by aridity, when the streams were very
heavily loaded with sediment. Consequently the flood plains were built
up from 10 to 30 feet as the case might be, and the evidences of former

THE FLUCTUATING CLIMATE OF NORTH AMERICA. 393
human occupation were completely covered. Finally, another slight
change, which we infer to have been in the direction of moister conditions,
has led to the renewal of dissection by the streams. The later phases of
this last process are complicated by the coming of the white man and by
the over-grazing of the hillsides by his cattle with the consequent tendency
for the soil to be washed away. Hence the evidence of the terraces needs
to be searched with much care. Nevertheless, it appears quite certain that
before the recent settlement of the country by Europeans, and after the
days of the primitive Hohokam, there was an interval of considerable
length, during which there occurred at least one climatic cycle involving a
change from moist to dry conditions and possibly back in some slight
measure to moist.
Leaving the somewhat inconclusive terraces, let us turn to the works of
man. A broad view of the ruins of the south-west seems to show that they
belong to at least three periods. During each of these periods the area
capable of cultivation appears to have been more extensive than at present,
although only slightly so in the last period. In the middle period the
arable area was larger than in the last, and in the earliest still larger.
The last period dates back only to Spanish times. Evidence of it has as
yet been discovered in only a few places, but, nevertheless, it seems con
clusive in itself, and still more so when taken in connection with other
phenomena. The best example for discussion here is the ruins of Gran
Quivira in central New Mexico. These ruins lie on the top of a rounded
hill about 200 feet above a broad open valley draining toward the south
and a mile or more in width. The altitude of the region is over 6000 feet,
so that the temperature is comparatively low. The ruins consist of two
distinct portions, Pueblo and Spanish, covering an area about 700 by 350
feet. All the buildings were constructed of light grey limestone broken
into rough rectangles. The exact source of the material is not evident,
although stone of this sort is visible in small outcrops not far away. Its
character, however, is such that long and arduous work would be required
to get it out in large quantities without the aid of explosives. Inasmuch
as the village was evidently built long before the coming of the Spaniards,
we must assume that the Pueblos, or their predecessors, put themselves to
a vast amount of labour in the process of quarrying, squaring, and trans
porting the stones of their numerous houses. Many of the dwellings were
manifestly of two stories, and the height of the heaps of rocks makes it
highly probable that some were of at least three stories. The rooms are
all small, as is customary in this region, the majority not exceeding 7 by
9 feet. The exact number of rooms has never been counted, but if we
assume that only half of the five and a half acres covered by the rnins was
actually built upon, and that the rooms, including the walls, had an
average size of 10 feet by 10, there must have been about 1100 rooms on •
the ground floor. The upper stories may be put at 400 rooms, although
the actual number was probably two or three times as great. This gives

394 THE FLUCTUATING CLIMATE OF NORTH AMERICA.
1500 rooms as a low estimate, which would mean at least a thousand
people. When the Spaniards came to the country at the beginning of the
seventeenth century, the village of Gran Quivira was evidently one of the
most important in the district. Otherwise the canny Fathers would not
have built here one of their largest missions. Possibly a portion of the
village was in ruins, and the rocks from it may have supplied materials
for the large church with walls 5 feet thick, and for the other structures
which the Spaniards erected. Nevertheless, the number of natives must
have been considerable. The beginning of the Spanish regime here, as in
the rest of New Mexico, appears to have been highly peaceful and pros
perous. Its end, so far as Gran Quivira is concerned, seems to have come
at the time ofthe Pueblo rebellion which culminated in 1680. Since that
time the site has been left as a centre around which a multitude of tradi
tions have gathered. One tradition ascribes its destruction to an earth
quake, another to a flow of lava bursting forth some miles away, and still
a third speaks of a river which has now disappeared.
The truth seems to be that there is no village now at Gran Quivira
because there is no water and the land is too dry for successful cultivation
except in years of good rainfall. A ranch is located in the valley below
the ruins, but it is not permanently inhabited, although a little cultivation
is carried on. Settlers have recently taken up land 10 to 15 miles to the
north, but are having a very hard time. If the rainfall is propitious they
can exist, but in 1909 none of them raised enough to live on. It scarcely
need be added that all depend upon deep wells for water. So far as we
can gather, the Pueblo Indians, like their Hohokam predecessors, knew
nothing of lime or mortar, and had no facilities for making watertight
cisterns. They constructed numerous reservoirs, however, and these were
their main dependence in the dry season. One such reservoir still
remains intact at Gran Quivira. It lies about a quarter of a mile east of
the village, in the mouth of a shallow arroyo, as dry valleys are here
called. The reservoir is only about 75 feet in width and 5 feet deep.
The owners of the ranch down below in the main valley say that during
seven years of life here they have never seen any water in it except
immediately after rain. My visit took place in the early spring of 1911,
after a more than commonly rainy season. The day before I started from
the railroad at Willard, 30 miles to the north, there was a heavy
storm, and during the drive we were soaked in a pouring rain. Neverthe
less, the next morning the reservoir contained no water, and showed no
sign of having had more than a small pool in it the day before. In all
the region within a score of miles of Gran Quivira there is only one
permanent spring. That is located 7 miles to the west at Montezuma, and,
as might be expected, it has its own ancient village. Strangely enough,
however, the Montezuma village was evidently abandoned long before
Gran Quivira. This suggests that the difficulty of raising crops was a mor§

THE FLUCTUATING CLIMATE OF NORTH AMERICA. 395
serious matter than the difficulty of obtaining water. At Montezuma the
land does not lie so low and flat as at Gran Quivira, and is not flooded as
are the lowlands of the latter place during summers when the rainfall is
unusually large.
In addition to Gran Quivira another similar ruin of a Spanish mission
deserves mention in order to show that the phenomena just described are
not isolated. This is the ruin of Buzani, about 12 miles below Caborca
on the lower Altar river, in north-western Mexico. Caborca is the last
inhabited place on the river. Farther downstream there is no water except
during the brief season of floods. At- Buzani a few Papago Indians
cultivate a considerable quantity of land in good years, but do not live there
all the time. They might remain through the year, for they have a well ;
but it is very deep, and the labour of drawing water is great. Here, as in
the other case, the Spaniards established a mission in a place which no
sensible man would choose for the purpose. I have not been able to ascer
tain the date of the Buzani church, and am not sure whether it dates from
the seventeenth or eighteenth century, but probably from the latter. Its
evidence is by no means so clear or pronounced as that of Gran Quivira.
In both these cases the point to be borne in mind is this : we have
before us two theories which stand on an absolutely equal footing as to
innate probability. The only question is which one best fits all the facts.
One theory is that the climate of the past three centuries has been
uniform ; while the other is that the seventeenth century was con
siderably moister than the nineteenth, and that the eighteenth century was
intermediate between the other two. Viewing the two theories without
prejudice, the theory of change seems to fit the facts better than the
theory of uniformity. Hence we may tentatively conclude that the seven
teenth century was a period of relative humidity, although both the preceding
and succeeding centuries may have been comparatively dry.
The other two periods of relative moisture may well be discussed
together. One of the most striking examples of an abundant population
in a region now almost uninhabited is the Chaco canyon in the north-western
corner of New Mexico. It lies just to the west of the wooded crest which forms
the continental divide in this portion of the great south-western plateau.
As soon as the wooded area is left behind one comes out upon a barren
treeless plateau at an elevation of 6000 or 7000 feet, a dreary region
where one may ride 25 miles at a stretch without seeing a single habitation.
The only inhabitants are a few nomadic Indians of the Navajo tribe
and a handful of white settlers. The Indians and a few of the white
men keep sheep, while the rest of the white men, which means about one
for every thousand square miles, are traders who supply the simple
wants of the Indians, and incidentally make large profits. Here and there
a few acres are cultivated by the Navajos, but not one family in ten possesses
any arable land, and even those who have fields by no means get their
main living or even a good share of it from agriculture.

396 THE FLUCTUATING CLIMATE OF NORTH AMERICA.
It scarcely need be said that the whole region is full of ruins. As one
of the traders put it after we had been talking about the extraordinary
number of ruins in one of the deep canyons which transect the plateau,
" I'd gamble that in any one of them canyons you'd find as many ruins
as you do here." The most remarkable of the canyons is that of
Chaco. In this steep-sided, flat-floored valley there are at the present time
two Indians who are reasonably sure of a crop of corn each year. I saw
their farms, unbelievably dreary wastes of drifting sand in the bottom of
the canyon where two large branches join, and where there is consequently
more water than anywhere else for 50 miles. The sand is a necessary
adjunct of farming, for it is needed to act as a mulch to prevent the
evaporation of the precious water. These two men between them have not
more than 20 or 30 acres, if as much. Even their crops fail completely
sometimes, as in 1903. During the last sixteen years, that is from 1895
to 1911, according to information given me by a trader's wife who has
lived there during that time and who moved away because her husband was
murdered by the Indians, the crop of all the Indians except the two just
mentioned failed absolutely in the years 1902, 1903, and 1904, while good
crops were raised in only two years, 1905 and 1908. During the remaining
eleven years, an Indian here and there raised a scanty crop, but none of them
could get a living were it not for their cattle and horses.
The contrast between the past and the present is remarkable. In the
Chaeo canyon or on the plateau on either side of it there are about twenty
ruins of considerable size within a distance of 25 miles. Some of these,
such as Pueblo Bonita, Chetwelketl and others, are strongly built, compact
structures, which must have sheltered hundreds of people, and the larger
ones probably had one or two thousand denizens. There can scarcely have
been less than 5000 people in the Canyon and its vicinity, and I should not
be surprised if the number were much larger. Whether it be 10,000 or 1000,
however, matters little, for the main point is that we have here a series of
strongly built, fortified villages whose inhabitants evidently cultivated a
large amount of land where now no crops can be raised. These people, as
appears from their pottery, their skulls, and their methods of architecture,
belong to a different civilization from that of the modern Pueblos who in
habited Gran Quivira at the time of the coming of the Spaniards. They had
evidently disappeared long before that date, as is evident from the present
ruins of their villages, and from the absence of any hint of their existence in
the early accounts of the country.
The builders of the high walls which we now see in the Chaco canyon
were not the original inhabitants of the country. Digging down below
their ruins one finds traces of an older occupation. Moreover, the largest
villages with houses of several stories, which are now in chief evidence,
invariably lie near to main lines of drainage or to places where it is clear
that a moderate increase in the amount of rainfall would cause permanent
streams or springs to flow. Farther away on the top of the plateau, however.

2® . J&

'¦'^4^^^

-itema- 3£$s±

RUINS IN CHACO CANYON.

RUINS IN CHACO CANYON.

YOUNG AND MIDDLE-AGED SPECIMENS OF SEQUOIA GIGANTEA.

A STUMP OF SEQUOIA GIGANTEA. A SECTION CUT FROM THIS STUMP WAS EXHIBITED
AT THE WORLD'S COLUMBIAN EXPOSITION IN CHICAGO IN 1893, AND IS NOW IN
THE AMERICAN MUSEUM OF NATUEAL HISTOEY IN NEW YORK CITY.

THE FLUCTUATING CLIMATE OF NORTH AMERICA. 397
ar trom the larger ruins and remote from any except small valleys, numerous
rums of another, more primitive type are found. They are usually small, and
are greatly ruined, and seem to belong to a time long anterior to the main
large ruins.
One of the best places for the study of this older type of ruins is in the
Pajaritan plateau 20 or 30 miles north-west of Santa Fe, the capital
of New Mexico. I was taken to this region by Mr. Kenneth M. Chapman,
curator and artist of the Archaeological Museum of New Mexico. Leaving
the stiff shada trees of the well- watered grassy plaza of mediaeval Santa Fe,
we took the narrow-gauge line of the Denver and Rio Grande railroad to
the lumber piles of the villageless station of Buckman on the Rio Grande
river. There we were met by Judge Abbott, who since his retirement from
the bench has made his home in the unique Canyon de Los Frijoles whither
we were bound. Crossing to the west side of the river, we left the barren
desert vegetation of yuccas, sage brush, and cacti which prevail in the river
valley at an altitude of 5500 feet, and climbed a thousand feet and more to
the plateau. The road winds up over variegated layers of volcanic tuff of
pale pink, yellow, or brilliant orange shades, interspersed with the darker
blue-black of basaltic lava-flows and capped with brick-red columnar tuffs.
These volcanic rocks form the Pajaritan plateau, a gently sloping, deep-
soiled region deeply cut by numerous canyons formed by streams running
eastward from the Jemez mountains. On the plateau we found a
beautiful district covered with forests of juniper and pinon, which at
higher altitudes give place to stately yellow pines set in open order with
stretches of sparse grass between them. The country was uncommonly
attractive as we drove slowly along, sometimes on the level, again dropping
into a hollow at the head of a canyon, and then climbing the slope once
more to the upland where we could look out east or west at great snowy
mountains. In spite of the deep soil, the grass and the trees, we saw no
sign of habitation, for except in a few insignificant spots in the bottoms of the
canyons where irrigation is possible, all the great plateau is too dry for any
cultivation below a level of about 8000 feet.
Soon after we had reached the main top of the plateau we came upon
the first of the great number of ruins which are scattered in every part of
the plateau. These particular ones were cliff dwellings of the usual type,
caves dug in the soft volcanic rock of the side of a shallow canyon, and
fronted by rooms made of blocks of the same soft tufa. The number of
such caves and cliff dwellings is literally thousands in this one Pajaritan
plateau. With them are associated the ruins of villages like those of the
Chaco canyon. After crossing several minor canyons we found ourselves
at the edge of the deep Canyon de Los Frijoles, or Bean Canyon, where a
precipitous cliff fell away 400 feet or more at our feet. Unhitching
the horses we left the wagon on the plateau, and turned the animals down
a steep winding trail, blasted in part from the face of the solid rock.
We had not followed them far when I uttered an involuntary exclamation of

398 THE FLUCTUATING CLIMATE OF NORTH AMERICA.
delight. I knew that we were to visit one of the most interesting ruins of
New Mexico, but I had no idea of finding it in so picturesque a canyon.
I had still less expectation of suddenly seeing far below us on a small level
space at the base of the precipice a structure which at first sight suggested
a Greek amphitheatre. It was the village of Tuyoni, excavated by the
School of American Archaeology at Santa Fe in the four seasons from 1908
to 1911. The plan of the ruins is symmetrical, a circle slightly flattened
on the north side, and containing from five to eight tiers of rooms arranged
like the seats of a theatre. Across the flattened end where the stage would
be expected, a line of rooms contains the remnants of three circular chambers
or "kivas," designed for religious purposes, apparently analogous to the
larger circular or elliptical structures which are found so commonly among
the ruins of adobe and wattle villages in the Santa Cruz valley and other
regions farther south.
The Canyon de Los Frijoles contains not only the main ruined village
of Tuyoni and several smaller ones, but also a great number of caves and
cliff dwellings. Doubtless the caves were at first the chief homes of the
aborigines; but as time went on and a higher stage of civilization was
reached, the excavations were used chiefly as store rooms, and the main
life of the households was conducted in rooms of stone plastered with
mud. Often a house consisted of three tiers of rooms in front of a cave ;
and in many cases the rooms were built one on top of another to a height
of three stories. Most of the rooms, like those of all the primitive people
of the south-west, as well as the modern Pueblos, were entered through the
roof. The small size of the rooms, not over 6 feet by 10 feet on an average,
is surprising. The reason, however, is clear. On the high Pajaritan
plateau the temperature often falls to ten degrees below zero Fahrenheit.
The relatively dense population must quickly have used up all tho
available dead firewood for many miles around, and it was no easy task for
a primitive people unsupplied with iron tools, to cut firewood sufficient for
anything more than the necessities of cooking. Farther south, or at lower
altitudes, the rooms were larger, for there it was easy to keep warm.
The low temperature does not appear to have diminished the number
of inhabitants. Frijoles canyon alone within a distance of not over a mile
and a half up and down the -narrow gorge, had a population of fully two
thousand souls according to the estimates of Dr. Edgar L. Hewett,
Director of the School of American Archaeology, who was personally in
charge of the excavations. The actual number of rooms, including the
village amphitheatre, the caves, and the cliff dwellings, appears to have
amounted to about three thousand. At the present time, according to Judge
Abbot, who owns all the valley except the ruins, the amount of land that
can be irrigated amounts to twenty-one acres. Manifestly the ancient
Pajaritans climbed out of the canyon to their daily work, and cultivated
the plateau where now not a solitary person can make a living from the
fruits of the earth.

THE FLUCTUATING CLIMATE OF NORTH AMERICA. 399
This leads us to a consideration of the older and more widely scattered
occupation of the country. During our drive to Frijoles canyon we watched
carefully not only for cave dwellings and villages of the Tuyoni type, but
for the location of little mounds which here and there at a distance from
the main sources of water, proclaim the location of houses scattered all over
the plateau. One who is not closely on the watch may miss these entirely,
for they are merely small heaps of stones. In the space of 7 miles we
saw houses of this type within sight of the road in forty-nine different
places. Inasmuch as several houses were often clustered in one group, the
total number of dwellings was sixty-seven. They were obviously mere
farmhouses, but some had from eight to twenty rooms, and must havo
been inhabited by more than one family. Therefore in our 7-mile drive
through the open, park-like forest we must have found within sight of the
road the dwellings of approximately a hundred families. It would be a
populous farming district in any part of England where one could find a
hundred families on 7 miles of road. We cannot, of course, assume that
absolutely every one of these houses was occupied at one time, but it
is not at all probable that any large number were vacant at a time when
new ones were being built. Well-squared blocks of stone must in those
days have been too valuable to permit of their being wasted when new
houses were to be built. Even in these days when metal tools and
explosives make it easy to quarry new stone, and when the population
is much less dense than formerly, the great ruins of antiquity in western
Asia are in constant danger of being utterly destroyed by the natives, who
carry away the stone for use in new houses. In the days of the' Pajaritans,
when the blocks of stone had to be hewn with stone axes and carried from
the quarries in the canyons on the backs of men, or rather of women, we
can scarcely believe that the people were so extraordinarily industrious or
so superstitious that they would leave good stones in ruins close at hand
and go to the labour of dressing new ones. Therefore we believe that at the
height of the prosperity of this region the site of practically every ruin was
occupied by an inhabited farmhouse.
These scattered little ruins, almost unnoticed even by the archaeologist,
present one of the most interesting problems in American archasology.
The potsherds found in them are of a different type from those found in the
larger villages or in the majority of the cliff-dwellings immediately around
them. The pottery of the farms, as Mr. Chapman points out, is almost
wholly a fine-grained ware painted white and adorned with geometrical
designs in black. In the larger, more modern ruins, however, only a little
of this is found, while the commonest kinds are a coarser white ware with
more abundant curves in the designs, and a wholly different type of red
ware adorned with black figures painted with a species of glaze. These
differences, coupled with other evidence such as the manifestly greater age
of the small isolated ruins, show that here, even more plainly than in the
Chaco region, we have to do with two occupations as distinct from one

400 THE FLUCTUATING CLIMATE OF NORTH AMERICA.
another and from the later and far less extensive Pueblo occupation of the
country, as are the modern American and Spanish occupations. The first
inhabitants spread far more widely than their successors. They seem to
have felt no need of being near the main sources of water nor yet of
gathering together as the later people did in places which could easily be
defended. For a long period before the advent of the enemy which finally
displaced them, their lives were free and comfortable in their high forest
homes. How or why they vanished is as unknown to us as is their origin,
but perchance we shall learn the story little by little. It will not be
a story of peace and monotony, for those are not the conditions which
prevail when a race comes into a country, nor when it is forced out. We
can scarcely doubt that stirring times took place — raids, plunder, repeated
invasions, great distress and the final disappearance of one type of civiliza
tion and its replacement by another. And this painful process of a change
of civilization took place not once alone, but at least twice. Formerly the
cliff-dwellers who built the compact villages like Tuyoni and Pueblo Alto
were supposed to have been of the same race as the modern Pueblo Indians,
but now we know that this is not true. Possibly, nay probably, the modern
Pueblo is related to the second or village-building type of ancient inhabi
tants, whom we may call Pajaritans in distinction from the still older type
who may perhaps be classed as Hohokam, but the relationship is not close.
The bones of the dead, exhumed after centuries, tell something of the tale.
The modern Pueblo Indian is braohycephalic according to Dr. Hrdlika ;
his head is relatively broad, as any one can tell by looking at him. Some,
however, are dolicocephalic, with long heads, but these are in a minority.
The present Indians are clearly of a mixed race. Their predecessors, on
the contrary, were ofa pure race, predominantly long-headed like ourselves.
Therefore we infer that they were conquered by invading broad-heads, and
that finally the invading broad-heads and as many of the long-heads as had
neither fled nor perished became amalgamated into a single race. Perhaps
the ancient farmers, the mediaeval villagers, and the modern Pueblo
Indians were not the only races which have passed across the stage of
history in the prehistoric days of America. In other parts of the south
west faint glimmerings are seen of still other cultures, which show that
change and movement have been as characteristic of the ancient history of
America as of that of Europe and Asia.
The dates of the three types of civilization of which we have found
evidence cannot well be determined. Dr. Hewett believes that the
traditions of the Pueblos may be relied upon as showing that it is at least
eight hundred years since the villages of the Tuyoni type were occupied,
but he attempts to assign no exact dates. One point, however, may here
be emphasized. In Asia, as I have shown in ' The Pulse of Asia ' and
more fully in ' Palestine and its Transformation,' we have evidence which
suggests that each of the chief dry epochs has been characterized by
great movements of peoples. The first such movement whose date is

THE FLUCTUATING CLIMATE OF NORTH AMERICA, 401
well defined was about 1200 b.o. At that time the ancestors of the
Greeks came into their peninsula, the Hebrews entered Palestine, the
Aramaeans from Arabia spread out into Babylonia and all the neigh
bouring lands, and Egypt was overwhelmed by invaders from both the
Libyan and Arabian deserts. The next great period of aridity culminated
in the seventh century or thereabouts. Its approach was marked by the
barbarian invasions of Europe, and its culmination by the Mohammedan
outpouring from Arabia. Finally, the third of the more important dry
epochs reached its climax in the thirteenth century a.d., when the
hordes of Ghenghis Khan ravaged Asia from China to the Mediter
ranean. Beside these more intense periods of aridity there seem to have
been others of minor importance, but these need not here be con
sidered. The history of Asia during the last three thousand years thus
consists of three main epochs of moist climate and advancing civilization,
separated by epochs of aridity and declining civilization accompanied by
migrations and wars. In comparing America with Asia it is certainly
striking to find that in the drier portions of the western hemisphere
we also have three main periods of prosperity and apparently of abundant
precipitation separated by epochs of decline and depopulation. Perhaps
the ancient farming population of America, the Hohokam, may date from
the period of moist climatic conditions at the time of Christ and earlier.
Their disappearance may have been due to the aridity of the period which
culminated in the seventh or eighth century. Then the village people, the
Pajaritans, may have flourished in the middle ages, and may have been
ousted by the twofold disaster of prolonged drought and fierce invasion
which would have come to America about 1200 a.d. if conditions here were
like those of Asia. And finally, the occupation of places like Gran Quivira
by the modern Pueblo Indians may have been made possible by the
propitious conditions of relatively abundant rainfall which followed the
aridity of the thirteenth century. I do not advance this as more than a
working hypothesis, but as such it may suggest various lines of research
hitherto neglected. Part III. — The Evidence of the Trees.
The evidence of climatic changes thus far presented probably seems
much more conclusive to the writer than to the reader. I am well aware
that when the name of an author once becomes identified with a theory his
fellow-workers are apt to discount his results. They know that a man
is prone to find what he looks for, and that when certain facts are capable
of two or more equally plausible- explanations he is likely to choose in
accordance with his preconceived ideas rather than according to the weight
of the evidence. In searching for some method wherebyjthis danger might
be avoided, I found no success until an article by Prof. A. E. Douglass of
the University of Arizona appeared in the Monthly Weather Review for
June, 1909, under the title " Weather Cycles in the Growth of Big Trees,"

402 THE FLUCTUATING CLIMATE OF NORTH AMERICA.
Prof. Douglass there shows that the thickness of the rings of wood
formed each year varies so closely in harmony with climatic conditions
that it may be used as a measure of the climate of the past. Taking about
twenty sections of trees from two to five hundred years of age, he made
micrometer readings of the thickness of each individual ring, and thus
obtained the average growth for each year. Most of his trees were yellow
pines which grew in northern Arizona at an altitude of about 7000 feet in
an open forest like that of the Pajaritan plateau already described. At this
altitude the seasons are sharply defined, the mean temperature for January
being 29° and for July 65°. Each annual ring consists of two parts. The
inner portion is relatively broad, soft and white ; its cells are round and
well fed ; and its period of growth is the early part of the growing season
when the soil is still well supplied with moisture from the snows of the
preceding winter and when the temperature is favourable to rapid growth.
The outer portion of each annual ring is narrow, hard, red, and pitchy ;
its cells are lean and emaciated ; and its period of growth is the later part
of the growing season when the soil is dry or when the temperature falls
low on the approach of winter. Occasionally a double ring is formed, but
this is rare and can usually be detected. Except where low temperatures
or extreme humidity prevail, the thickness of the rings is, on the average,
proportional to the rainfall, for the amount of growth each year depends
upon the available supply of food, which in turn is determined by the
amount of moisture. This is true more or less in all regions, but
especially in places like Arizona where the temperature of summer
is high enough to promote rapid growth, while the water-supply is
limited, and the life struggle of the tree is against drought and not
against cold nor undue moisture nor against its fellows or members of
other species. The measurements made by Douglass show that the
theoretical expectation is realized, for the curves of growth and of rainfall
for the forty years or so since records are available present a marked
agreement. Manifestly, then, the curves of growth may be used as a means
of determining the rainfall in periods long antecedent to any human
records. Douglass has so used them, and his analysis indicates cycles of
precipitation with a periodicity of 11, 21-2, and 32-8 years, a result which
agrees with the cycles indicated by records of rainfall. He did not attempt
to investigate larger cycles, nor to determine the relation of the climate of
periods separated by hundreds of years. Obviously this is possible, and it
needs only an expansion of Douglass' method, and a study of the sources of
possible error to enable us to come to definite conclusions as to the nature
of the climate at any time back to the youth of the oldest available trees.
During the year 1911 I began to make use of the method of Prof.
Douglass and obtained some most interesting results. In the first place,
Prof. H. .S. Graves, Forester of the United States Forest Service, most
courteously placed at my disposal several thousand stem analyses obtained
from old trees by the members of his Bureau in various parts of the Unite4

THE FLUCTUATING CLIMATE OF NORTH AMERICA. 403
States. In the second place, I went to California on behalf of the Carnegie
Institution, and with the help of two assistants made similar analyses of about
two hundred of the " Big Trees " or Sequoia ijigantea of the Sierra Nevadas.
The term " stem analysis," as used by foresters, means the measurement of
the thickness of the rings of annual growth beginning at the outside and
working inward. When the year in which the tree is cut is known, it is,
of course, an easy matter to determino the rate of increase in the diameter
of the trunk for any year desired during the life of the tree. Stem
analyses are made by the Forest Service at various points along the length
of felled logs while they are in process of being sawed for the market.
Only those at the base, however, that is " stump analyses," have been used
in the present calculations, as they appear to give the most accurate
results, and as they are the only kind possible in the case of the Big Trees
of California. The measurements of the Forest Service and my own were
made on the basis of ten-year rather than annual periods. That is, the
thickness of the rings was measured in groups of ten instead of for in
dividual years. Thus we can estimate the rainfall for the decade 1801-
1810, but not for the specific years 1803, 1809, and so forth. In the
present paper I shall take only two samples from among the considerable
number of species of trees whose analyses Prof. Graves placed at my dis
posal. The others will be presented later in the final report already men
tioned. Also, I shall not attempt to discuss in detail the methods of
correction which have been employed in order to eliminate the effect of the
decrease in the rate of growth which is the normal effect of increasing age.
I shall leave that, together with certain other more mathematical considera
tions for an article in the American Journal of Science and for the report on
" The Climatic Factor."
Let us first ascertain what may be learned from a study of the trees of
New Mexico which grow in the mountains close to the ruins which we
have hitherto been considering. The courtesies of the Forest Service have
included not only the placing of a large number of stump analyses in my
hands by the chief of the Bureau, but also some most helpful compilations
by other members of the Service. I am especially indebted to Mr. A. B.
Recknagel, Chief of Silviculture at Albuquerque, New Mexico, who has
gathered a most valuable series of statistics for the yellow pines of that
state, and has most generously allowed me to use them. His data are
based on 645 trees from four of the United States Forest Reserves, as
follows : — Gila forest. Jemez' forest. Datil forest. Zuni forest.
Over 200 years of age ... 53 trees 10 trees 74 trees 164 trees
Under „ ,, ... 124 „ 19 „ 66 „ 135 „
Total number of trees over 200 years of age  301
„ „ „ under „ ,  344
Grand total  645
The Gila and Datil National forests are in the south-western part of

404 THE FLUCTUATING CLIMATE OF NORTH AMERICA.
New Mexico near the Arizona line. They lie between the ruins of the
Santa Cruz valley and those of the Jarilla mountains, and about equally
distant from both. The Jemez forest lies in the centre of the northern
part of New Mexico, and adjoins the Pajaritan plateau where the ruins of
Tuyoni and the Canyon de los Frijoles are situated. The other reserve,
the Zuni forest, is situated in north-western New Mexico somewhat more
than 50 miles south of the remarkable ruins of the Chaco canyon. Thus
the forests are scattered in such a way that they have approximately the
same distribution as the ruins with which we have been dealing. In general,
however, they lie at greater ;altitudes than the ruins. " The measure
ments," to quote Mr. Recknagel, " were all taken within the Western
Yellow Pine type between altitudes of 7000 and 9000 feet. They were taken
in the main body of the Western Yellow Pine type, and, therefore, can be
considered as belonging to members of the pure stand of the species."
In other words, the trees with which we are concerned grew in the
portion of the yellow pine area where the trees grow best and where they
are neither at the lower limit so as to be especially liable to injury by
drought nor at the upper limit so as to be especially liable to injury by
excessively low temperature or long winters. Moreover, they were not
selected with a view to any special characteristic other than age. All the
available analyses of trees over 200 years of age were used, and to them
were added an approximately equal number of younger trees in order to
afford fair comparisons.
From Mr. Recknagel's figures I have computed how fast the average
tree grows during the first, second, and third decade, .and so on to
the end. This shows that the average tree grows 0-46 inch during the
first decade of its life, 0'63 during the third, 0-40 in the twelfth, and only
0-23 in the thirtieth. Obviously if we are to compare the climate of
different periods during the life of the tree, we must eliminate the
differences due merely to the fact that the tree has grown older. For
instance, a growth of 0-46 in the first decade indicates no more propitious
conditions of climate than a growth of 0-23 during the thirtieth decade
when the tree is 300 years old. All that we need to do is to multiply the
rate of growth of the old trees by the proper factor, two in this case, and
the result will show the relative effects of climate. For example, suppose
that in the decade 1610-1620 a tree grew 0-40 inch, and in the decade
1901-1910 it grew 0-19 inch. Multiplying the growth during the last
period by two, which is the corrective factor for the thirtieth decade, we
get 0-38 inch as the rate which the tree would have grown had it been in
its first instead of its thirtieth decade. The difference between 0'40 and
0-38 is due to some external cause. The external cause may be an
accident, but when a large number of trees are averaged together, the
influence of accidents disappears and the remaining differences are attri
butable to climate. One more correction, however, must needs be applied.
A study of Mr. Recknagel's figures, as well as of many others, shows

THE FLUCTUATING CLIMATE OF NORTH AMERICA.

405

clearly that if a tree is to live to old age, it cannot grow fast in its youth.
That is, if we take a large number of trees so as to obtain averages, the
tree which grows faster than the normal rate in its youth dies at less than
the average age, while one which grows abnormally slowly stands a good
chance of living to an exceptional age, provided the slow growth is not due
to disease. Hence, before we can compare the rates of growth of different
trees, we must apply another correction which I have called the correction
for longevity. This applies to the old trees during their youth, but not in
their old age. Hence our figures 0-40 must be multiplied by another
factor giving 0-41 as the final result for comparison with 0'38. These two

tnchesv, 0-6

O OOOOOOO OOOOOOOOOOCiOOOOO ooooooo

OS

0-4

0-3

0 2 —

O I

f
t
\
1 \ % \ \
k
1
t
1
t
1
/
1
$
1
t
\ \ \
\
\
I
1
1
1
1
1
1
1
t
*,
*<.
*x
V,
N
, .
A
\
1
1
1
/*¦
s
?
?
*>**
«*.
^"n
t
f/
/'
-<*
"V
t
V
---
\%
¦•«.
.
/'
^
*^
i
9.
1
^
N,.
 -.  Uncorrected Curve.  Curve corrected forage.
Curve corrected for age and longevity
FIG. 1.— CDRVE OF GROWTH OF 50 YELLOW PINES OVER 280 YEARS OF AGE.
types of correction sound somewhat complicated, and so. they are. The
point to be noted, however, is that they are purely mathematical, and that
they do not depend upon the individual judgment of the investigator.
While he is working out the calculations he cannot have any idea of the
exact results which will be obtained ; nor can he make the results fit one
theory rather than another except by wilful manipulation.
The accompanying diagrams, Figs. 1 and 2, show the results of Mr.
Recknagel's tree analyses when computed according to the method here
described. In both figures the horizontal line indicates the course of time
from 1590 to 1910 a.d. The height of the curves indicates the relative
406

THE FLUCTUATING CLIMATE OF NORTH AMERICA.

rate of growth of the trees at any particular date. Fig. 1 is given simply
as an example to show the effect of the corrections. It is based upon the
fifty oldest trees, which began growth at periods ranging from 1460 to
1590 a.d. The upper dotted line indicates the actual average rate of
growth of these fifty trees without any correction whatever. The lower
dash line is the same curve after the correction for age has been applied,
that is, after allowance has been made for the fact that young trees grow
faster than old. The solid line shows what happens to the curve when the
correction for longevity is added to that for age ; that is when allowance
is also made for the fact that during their youth trees which are to live to
a great age regularly grow more slowly than the normal rate.
In Fig. 2 the solid line represents the curve of growth and presumably
of climate as derived from the 272 of Recknagel's trees which began to
grow before 1700 a.d. The numbers in parenthesis at the bottom indicate
the number of trees on which the curve is based in each decade. From
1700 a.d. onward, the number remains uniform, since no new trees are
added. The dotted line is the curve of growth of the red fir in the Lake

•^^ OOOOOGOOOOGOOOOOOOCJOCiOOOOOOOOOOO o-^-
<- ;SJ ti.o(o«)«)(o»oio(oiD(otsrs(sf^rsi^fv(^p^r>cDCO(Dto(Da)foai(Daioi ojr/^cg

FIG. 2. — CURVES OF GROWTH OF YELLOW PINE (SOLID LINE) AND BED FIR
(DOTTED line).

Tahoe National forest in California. At first sight this curve seems to
be almost diametrically opposed to that of the yellow pines of New Mexico,
but there is no real discrepancy. In fact, the seeming disagreement is,
perhaps, the strongest possible confirmation. The red fir of Tahoe grows
at an altitude of about 8000 feet in latitude 39° near the crest of the
Sierra Nevada mountains, 250 miles north-east of San Francisco.
Although the altitude is the same as that of the yellow pines of New
Mexico, and the latitude only 4° or 5° farther north, climatic conditions are
quite different. The precipitation is much greater than in New Mexico,
and comes largely in the form of snow. Moreover, it continues through
the spring, which season is dry in New Mexico. The result is that in the
mountains above Lake Tahoe the snow in ordinary years remains upon the
ground well into June or even July, and the ground is saturated with
moisture practically all the time. Too much moisture, and especially too
long a continuance of the snow, are the factors which retard the growth of
the red fir, whereas too little moisture and short, open winters are detri
mental to the yellow pine. Hence we should expect that climatic condi
tions which would favour the pine, that is long, snowy winters, followed by

THE FLUCTUATING CLIMATE OF NORTH AMERICA. 407
abundant precipitation in the spring, would check the growth of the fir.
In general, that is what we find ; the two curves are almost the reverse of
one another, a condition which does much to substantiate the idea that the
climate of the past may be read from the trees.
Turning now to the interpretation of the New Mexican curve, we find a
surprisingly close agreement with our conclusions based on the evidence
of terraces, archaeology, and history. Before going on to discuss the matter
I would emphasize the fact that these conclusions were all reached before
the trees had been investigated. They have been set down above exactly
as they were reached, at a period of from fifteen to five months before
any examination of the trees was made. The importance of this lies in the
fact that the agreement of the mathematically derived tree curves with the
conclusions derived from entirely different methods furnishes the strongest
possible confirmation of the accuracy of those methods as employed both
in America, Asia, and southern Europe. From the ruins of Gran Quivira,
it will be remembered, we concluded that at the time of the Spanish occu
pation of New Mexico in the first half of the seventeenth century climatic
conditions were distinctly more favourable than at present. About 1680
the ruins of Gran Quivira seem to have been finally abandoned at the time
of the Pueblo rebellion. During the succeeding century conditions
appear to have been somewhat better than during the one that ended in
1900, as appears probable from the ruins of Buzani, and one or two other
places not here mentioned. The curve of the trees points to exactly the
same conclusion. One reason why the Spaniards were able to establish
themselves in New Mexico almost without a blow may have been that from
1600 to 1645 the amount of rain and the general conditions controlling
the growth of vegetation were more and more favourable. Places like
Gran Quivira were readily habitable and were growingly prosperous so
far as their prosperity depended upon good crops. Under such circum
stances the ignorant Pueblos would naturally look upon the coming of the
Spaniards as a blessing. A decrease in rainfall begins to be apparent
about 1645, but it is unimportant at first because the amount of growth of
the trees, and impliedly of the crops, continues to be well above the normal
until about 1665. Thereafter a rapid deterioration takes place, which cul
minates about 1680 or soon after. At exactly that time a widespread
uprising took place against the Spaniards, the Pueblo rebellion, which
ousted the Europeans for some years. Famine must have prevailed as
the climate became drier, and the distress occasioned would almost surely
lead the Indians to ascribe their misfortunes to their conquerors. Curiously
enough a document has just come to light which confirms our conclusion
as to famine. I am indebted to Mr. E. E. Free, of the Bureau of Soils of
the U.S. Department of Agriculture, for calling my attention to an account
of an old census given by Mr. J. W. Curd, in the El Paso Times. El Paso,
although in Texas, lies almost at the middle of the southern boundary of
New Mexico, The census is dated September 11, 1684, and contains a list

408 THE FLUCTUATING CLIMATE OF NORTH AMERICA.
of 109 Spanish families living near El Paso. It is signed by the Spanish
governor. "While the document is nothing more than a dry and un
interesting census roll," to quote Mr. Curd, " it is illuminative of the
terrible devastation and suffering that resulted from the Indian revolt in
1680. This revolt destroyed some forty-two presidios and missions in
New Mexico north of El Paso, and the remnant of Spaniards and friendly
Indians took refuge in Guadalupe del Passo (El Paso). While the
Mission Guadalupe was a rich one, and additional supplies were forwarded
from Mexico City, the people still suffered from lack of clothing and food.
The census shows that what crops were planted that year consisted only of
maize, which, owing to drought, was an almost total failure. What maize
was grown was eaten green, so there was no supply for winter." It is not
here possible to carry the matter further. The curve speaks for itself.
After 1680, conditions appear to have improved; the eighteenth century
was a period of less rainfall than during the time of the early Spanish
occupation, but of more than the last century.
The Sequoia gigantea of California is the largest and probably the oldest
known tree. During the summer of 1911 I was able to make measurements
of two hundred stumps which ranged from 6 to 25 feet in diameter, and
from 250 to 3150 years in age. Three were over 3000 years old, and
forty began to grow 2000 or more years ago. During 1912 a larger
number were measured, but it has not been possible to complete the
long tabulation and calculation which they require in time to use them
in this article. The sequoia grows in a region which is climatically
similar to New Mexico, except that it is colder and has no rainy season
in summer. Long, snowy winters and rains continuing well into the
dry summer are the conditions which favour growth. The curve of
the sequoia has been computed in the same way as that of the other
trees. The period covered by the growth of the trees, however, is so
great that the number of trees whose analyses are now available is not
enough to give certainty to the corrections. Accordingly, after obtaining
a curve corrected as accurately as possible both for age and longevity, I
averaged that curve with the original uncorrected curve in order to avoid
the risk of over-correction. This means that the curve here given, the
solid line of Fig. 3, is still subject to considerable correctipn. Nevertheless,
I think it worth while to publish it at this time with the warning that it
must not be accepted as final. The final curve, which will appear in the
report to be published by the Carnegie Institution, will resemble the
present one in its sinuosity, but the indications are that some of the maxima
and minima will be pushed a few decades earlier, and that the portion
from 1200 a.d. onward will not rise so high as in the present case.
Before we proceed to discuss the curve a word or two of explanation
may be added. The solid horizontal^ line marked 10 is added merely for
reference, to aid the eye in estimating the fluctuations of the curve. The
figures arranged vertically on the margin indicate the growth in millimeters

THE FLUCTUATING CLIMATE OF NORTH AMERICA. 409
per decade. The small underlined numbers in the body of the diagram
near the bottom at intervals of two hundred and fifty years indicate the
number of trees on which the curve is based. The dates in parentheses
at the bottom are placed there to show that there is a possibility of error
which will be corrected by further investigation. The extreme fluctua
tions of the early part of the curve give an exaggerated idea of tho
variability of the climate at that time. This is due to the small number of
trees available. When more have been measured the extremely sharp
character of the depressions and hollows, or of the arses and theses, as 1
have elsewhere called them, will disappear.
Examination of the curve shows that the climate of the interior of
California has been subject to marked pulsations during the past three
thousand years. For instance, at the time of Christ the tree grew 40
per cent, faster than at the end of the fifteenth century. Practically all the
trees at the latter date were of large size with thoroughly developed root
systems, and with a vast supply of strength stored up from the past.
Moreover, they were growing high among the mountains where the supply
of rain and snow was largest, and many of them were standing in swamps
or beside brooks. If they could be so affected by drought as to show a
decrease of 40 per cent, in the amount of growth, other less favoured
plants must have suffered much worse. It is noticeable that in general the
trees recover rapidly after a period of aridity and then fall off more slowly
as another dry time approaches. This is important as an indication that
climatic changes from dry to moist take place rapidly, while those in the
reverse direction are slow. It bears also on another point. It may be
suggested that the inequalities in the curve are due to accidents such as
fires. In the first place, this is not probable since the two hundred trees
were located in four different areas with a distance of 60 miles between
the extremes, and with high mountains and deep valleys intervening. The
rapid rise and gradual fall of the curves, however, disproves any such
supposition. An accident, such as a fire or anything else, would suddenly
cause the trees to grow slowly, after which they would gradually recover,
whereas the actual case is the reverse of this.
In Fig. 3 I have added a dotted line. This is the curve of climatic
pulsations in Asia as given in ' Palestine and its Transformation.' The
two curves disagree in certain places, but on the whole they are in surprising
harmony. The disagreements may be due to the absence of data in com
piling the Asiatic curve; for instance, between 1200 and 1000 b.c. I had
no data whatever, and hence merely drew a straight line. In other cases
the fact that indications of aridity happened to be especially well preserved
at a certain time, such as the seventh century of our era, may have caused
me to carry the Asiatic curve lower than was justifiable. It should be
noted, however, that in the fully corrected sequoia curve the lowest point
of all will apparently fall earlier than 720 a.d., which will agree with the
Asiatic curve more closely than is here apparent. Moreover, the longest

410

THE FLUCTUATING CLIMATE OF NORTH AMERICA.

continuous decline anywhere apparent in the California curve is from the
time of Christ to the beginning of the eighth century. The greatest dis
agreement between the two curves is found about 300 a.d. Whether there
actually was disagreement at that time, or whether I have made a mistake
in the Asiatic curve, I shall not here attempt to discuss. In general it
may be said that the three noticeable depressions in the Asiatic curve,
namely 300, 650, and 1200 a.d., are possibly all exaggerated because
special events due apparently to increasing aridity happened to culminate
at those particular epochs.
In spite of certain disagreements, by far the most noticeable fact about
the curves is their close agreement. Take the epoch centring at the time
of Christ, for example, or those which centre 1000 or 1600 a.d. The
agreement is so close that it cannot possibly be a matter of chance. Further

L. "

>
I
.'-
WX^
~x
H -
(
1 ~-
V
MX
A'
h
\
r\
¦\
J
V
U
1/
^W
'V
\r
M7
J
\
v
V
^
A
h
/
V
J-\
^
V
ffy
p.,\
<sr
\.
1 *•
V
rJ
V
w~-
J
1
i I
'.
/
!
l-
,-¦
-,
J
\
¦'
/
.¦-:¦
/ 1
i 5
/
/
25
f
7
76
/
16
•J
1.10
h
fl
im
i
?
If
fl
1 1
1
1
,3008 C.
(1390)
,000
(1080)
500
(5 SO)
S.CrAD. (30.B.C)
500
(480)
1000
(990)
FIG. 3. — CURVE OF GROWTH OF THE SEQUOIA GIGANTEA OF CALIFORNIA (SOLID
LINE), AND CHANGES OF CLIMATE IN ASIA (DOTTED LINE).
discussion of the subject must be deferred for the present. We can here
merely sum up the main conclusions. The study of the trees of New
Mexico and California in the first place confirms the conclusions derived
from examination of the ruins and physiographic evidences found in the
drier parts of North America. It thus shows that the methods upon which
those conclusions are based are sound, and that the results derived from
such methods whether in America or Asia are valid. In the second place
the trees confirm the theory of pulsatory climatic changes. They show
that the climate of the Earth is subject to pulsations having a period
of hundreds of years. In the third place the rate of growth of the trees
indicates that in the distant historic past the muist epochs were on the
whole moister than the similar epochs in more recent times. Fourth and
last we are led to conclude that the main climatic changes of America are
synchronous with those of Asia and are of the same kind. This does not
THE FLUCTUATING CLIMATE OF NORTH AMERICA. 411
mean that changes in tropical countries are like those in the temperate
zone. It does indicate, however, that in the temperate continental regions
of the world periods of exceptional aridity or of exceptional moisture have
occurred at the same time, and have sometimes lasted for centuries.
Hitherto this point has been open to question, and therefore historians and
other students of man have been sceptical as to the possibility that
climatic changes could have been of sufficient importance profoundly to
influence history. With the fuller application of the methods here dis
cussed, we shall soon be able to determine the exact nature and degree of
climatic changes throughout historic time, and then we shall have the
basis for a true appreciation of their effect upon history.

PRINTED BY WILLIAM CLOWES AND SONS, LIMITED, LONDON AND BECCLES.

3 9002 01497