injuring the beams in any way. Such an instrument was developed in the tubular borer as already described. This tool was sent to Mr. Morris and during 1920 he bored into all the beams at Aztec then available and sent me the cores.

These cores, together with other sections of beams too frail for boring, finally represented 37 different beams in some 20 different rooms scattered along the larger north part of the ruin. Practically all of these show similar rings near the outside, and by counting to the last growth ring of each it was easy to tell the relative dates at which the various timbers were cut.

In order to help in describing given rings in these various sections, a purely imaginary date was assumed for a certain rather large ring which appeared in all the timbers. This was called R. D. (Relative Date) 500, and all other rings earlier or later are designated by this system of relative dates. Many interesting results were evident as soon as the various relative dates were compared. In the first place, instead of requiring many hundreds of years in construction as any one would suppose in looking at the ruin, the larger part of it was evidently erected in the course of ten years, for the dates of cutting the timbers found in the large north side include only eight or nine years. The earliest timbers cut were in the northeast part of the structure. The later timbers are at the northwest, and it is evident that the sequence of building was from the easterly side to the westerly side, ending up with the westerly end and extending toward the south.

In one place beams from three stories, one over the other, were obtained. The top and bottom ceiling timbers were cut one year later than those of the middle ceiling, showing that in vertical construction the three floors were erected in immediate succession. A floor pole from Pueblo Bonito was cut one year later than the latest beam obtained from that ruin.

An even more interesting fact was soon after disclosed. A study of the art and industries of neighboring ruins had satisfied Mr. Nelson and Mr. Morris of the American Museum that some of the ruins in Chaco Canyon, some 50 miles to the south, were not far different in age from those at Aztec. The only beams immediately available from the Chaco Canyon ruins had been collected in the Pueblo Bonito ruin 25 years before by the Hyde expedition. Accordingly sections were cut from seven beams which this expedition had brought back to New York City. One of these sections was a cedar and has not yet been interpreted, but the others were immediately identified in age both among themselves and with reference to the Aztec timbers. It was found that these Pueblo Bonito beams were cut within a few years of each other at a time preceding the

1

cutting of the timbers at Aztec by 40 to 45 years. Many of the timbers of each ruin were living trees together for more than one hundred years and some even for two hundred years, and there seems no possible doubt of the relative age here determined. This result showing that a Chaco Canyon ruin was built nearly a half century before Aztec is the first actual determination of such a difference in exact years. A single beam from Peñasco, some 14 miles down the Chaco Canyon from Pueblo Bonito showed that its building was intermediate between Pueblo Bonito and Aztec.

Another association of growth rings with prehistoric deposits has rapidly developed in the last two years. In 1904 the writer discovered an Indian burial at a depth of eight feet in a cultivated field near Flagstaff, Arizona. A skeleton and two nests of pottery were revealed by a deep cut which a stream of water had made through the land. Near the burial was an ancient pine stump standing in place 16 feet underground. The tree was later discovered by a neighbor and became part of a bridge support. The Indian remains were given away except a red bowl of simple pattern and a good piece of black and white ware which is now in the Arizona State Museum. In 1920 the search for these buried trees was resumed and more than a half dozen in excellent preservation were found at depths from four to twelve feet. Mr. L. F. Brady of the Evans School gave most important help in getting out sections of these. In the summer of 1921 he again resumed the search and found several more buried trees and especially determined several levels at which pottery and other Indian remains are plentiful. These buried trees have been preserved by their pitch and show here and there quantities of beautiful little white needleshaped crystals, which Dr. Guild has discovered to be a new mineral and to which he has given the name "Flagstaffite."

Several conclusions are already evident in the study of these buried trees. In the first place they supply much desired material from which some data regarding past climates may be obtained. The trees buried most deeply have very large rings and a certain kind of slow surging in ring size. Both of these features are characteristic of wet climates. The stumps at higher levels show characters common in dry climates, that is, general small rings and a certain snappy irregularity with frequent surprises as to size. This variation with depth gives a strong intimation of climatic change. The cycles dominant at these different levels also may be read from these sections and are likely to prove of great value.

In the second place this material will help in determining the age of the Indian remains and perhaps even of the valley filling in which these objects were located. There are several ways of

getting at this which will take time in working out but there is one inference immediately evident. One log was buried only eighteen inches, yet its rings do not tally with the 500 years of well determined rings of modern trees in that neighborhood. Allowing about a century for the sap-wood lost from the buried tree and a half century more necessary to detect cross-identity, we have an approximate minimum of 350 years for that foot and a half of depth. The age of Indian relics at four and even nine feet must be very considerable.

These then are the first results of the application of the general study of tree rings to archeological work and suggest further possibilities. Not only does it seem probable that this beginning of relative chronology of the wonderful ruins of the Southwest will be extended to include other ruins in this region, but this study of the prehistoric writing in trees will help in the clearer understanding of the climatic conditions which existed in those earlier times when the largest bona fide residences in the world were being built.

IX. CONCLUSION

The economic value of this study of tree rings and climate is to be found in the possibility of long-range weather forecasting. In non-economic terms we are trying to get the inter-relationships between certain solar and terrestrial activities by the aid of historical writing in the trees. The work is not done; a wide door is open to the future. Hence it is impossible to make an artistic conclusion. There is no real conclusion yet. Some definite results have been reached and they encourage us to hope for larger returns in the future. Through this open door we can see attractive objectives looming above us and we note the outlines of some of the hills to be surmounted. To climb these metaphorical hills we need groups of trees from all parts of this country, from numerous specially selected spots and areas, from distant lands; we need ancient tree records from Pueblo ruins and modern Hopi buildings, from mummy case and viking ship, from peat-bog and brown-coal mine, from asphalt bed and lava burial and from all ancient geologic trees in wood and stone and coal. We need measuring instruments, workers, museum room for filing and displaying specimens. And we need great quantities of climatic data obtained with special reference to tree comparison. With all this and with a spirit behind it, we shall quickly read the story that is in the forest and which is already coming to us through the alphabet of living trees.

VARIABILITY VS. UNIFORMITY IN THE

TROPICS

By Professor STEPHEN SARGENT VISHER

INDIANA UNIVERSITY

IT

T is commonly stated that tropical climates are extremely uniform. This is only partly true. They indeed have comparatively slight seasonal variations in temperature and in the length of day and night, and large areas have rather steady winds much of the time. But continual emphasis upon the uniformity of tropical climates is misleading because there are important variations in temperature and wind, while the rainfall of the lower latitudes appears to be more variable on the average than the rainfall of higher latitudes. There likewise appears to be more variation in storminess and in rapid change of air pressure than in higher latitudes.

Recent field investigations (financed by the Bishop Museum of Honolulu and Yale and Indiana Universities) in Hawaii, Fiji, tropical Australia, the East Indies, the Philippines and tropical China, and examination of official meteorological records concerning these areas and others have disclosed interesting evidences of tropical variability and have convinced me that the conventional statements, based on averages, are sufficiently misleading to make it worth while to emphasize the climatic variability occurring in the tropics.

The small seasonal contrasts in temperature which are characteristic of the tropics are perhaps the chief reason why the impression has been gained that tropical climates are uniform in other respects. Another reason for this common belief is the fact that the climatic data concerning the tropics are chiefly available in the form of averages. Averages by themselves are very misleading and should be supplemented as soon as possible by statements as to extremes, and as to normal extent of departure from means.

Although the average seasonal range in temperature is indeed small in low latitudes as compared with middle latitudes, there is an appreciable seasonal contrast in latitudes more than 10° or 15° from the equator. Indeed some parts of the tropics have about as great a seasonal range of temperature as certain especially uniform parts of higher latitudes. This is illustrated when the average

differences in mean temperature between the three warmest months and the three coolest of the following pairs of seaport cities are compared. Some of the cities in the right-hand column are not within the tropics according to the narrowest limitation of that However, all are within the belt dominated by the Trades during most of the year, which is the belt commonly considered as tropical.

Even in regard to extremes of temperatures, some cities in fairly low latitudes have ranges which approach those of the less variable parts of the relatively high latitudes. This is illustrated by the following table showing the difference between the highest and lowest temperatures ever officially recorded at certain pairs of seaport cities up to a recent year.1

That there are appreciable seasonal contrasts in temperature in lower latitudes is not surprising when the seasonal variation in insolation is considered. In spite of the fact that the sun shines vertically somewhere between the two tropics every day in the year, there is a great change in angle of incidence. Few people realize that when the sun is vertically over the northern tropic

1 References to sources of data are given in a longer, more technical article on "Variability of Tropical Climates" to be published in The Geographical Journal (London).