Climatic Cycles and Tree-Growth/Chapter 3

by Andrew Ellicott Douglass

III. COLLECTION OF SECTIONS.

The material upon which the discussion of climatic cycles and tree-growth is based has been derived from 230 trees collected in the 15 years from 1904 to 1918. The regions drawn upon comprise chiefly Arizona with its yellow pine, the Baltic drainage area of north Europe with its Scotch pine, and the high Sierras of California with their great sequoia. Two small collections come respectively from the northeast and northwest coast of the United States. The collections have been made in small, convenient groups as opportunity offered, to each of which a name has been given which will appear below.

The relative dimensions of the various groups may be expressed in terms of the number of measures of rings. In the first Flagstaff group there were about 10,000. In the second Flagstaff group of 1911 only a few hundred. The Prescott groups included about 4,000; the 9 European groups about 9,000. The Vermont group had between 2,500 and 3,000, and the Oregon group about the same. The first collection of sequoias in 1915 had about 25,000 measures and the collection in 1918 embraced about 22,000.

Throughout the whole study it was desired to get as long records as possible and old trees were therefore selected. In nearly every case this meant large trees also. Apart from this no special selection of trees was made at any time, save only in the Christiania group, in which so many of the logs showed a "complacent" habit, with long successions of equal rings rather large in size, that some effort was there made to find the logs which showed variations in ring-size. A complacent ring-record without doubt means that the environment of the trees was well adapted for its best development.

THE FIRST FLAGSTAFF GROUP OF TWENTY-FIVE SECTIONS.

The plan of using tree-rings for the general purpose of a check on astronomical and meteorological phenomena was first formulated in 1901. The first measurements were made in January 1904, on a huge log in the yards of the Arizona Lumber and Timber Company at Flagstaff. This method of measuring was extremely inconvenient and the succeeding 5 sections were cut from logs and sent to town for more careful examination. Hence the exact location of these first 6 was never visited. The remaining 19 trees were selected in 1906 by myself in the forest while the logs were yet lying near their stumps, and I was able to mark on each section the points of the compass and otherwise describe the location. The measurements were completed in 1907 and published in the Monthly Weather Review of June 1909. They had not been subjected to cross-identification and, when the value of this process was recognized in 1911, the 19 sections of which samples had been preserved were compared and a complete cross-identification carried through. Thus the errors of identity in the former tabulation were found (published in 1914^[1]) and a complete new set of tables and averages made from the original measures. For a time it was thought that an error of one year might exist in the period of the great drought of 1820-23, but the various checks made upon identity lead easily to the belief that there are no errors of identity in this 500-year series.

Subgroups.—The trees of this group were divided into three subgroups consisting of (1) 6 trees from 3 miles south of Flagstaff; (2) 9 trees from 11 miles southwest of Flagstaff; (3) 10 trees from a point 1 mile west of the last subgroup. A comparison of the 3 subgroups clearly reveals the general character of the longer periods hereafter to be discussed and shows lesser variations to be common to all. Interesting differences, depending on the location in which the trees grew, have been mentioned.

First suggestion of cross-identification.—Other interesting facts came to light. It was especially noticeable that a given year of marked peculiarity could be identified in different trees with surprising ease. This is illustrated in plate 2, where shavings from 5 of the Flagstaff trees have been photographed; the photographs have been enlarged to such a scale that the distance from the large ring 1898 (indicated by the upper line of black crosses) to the small ring 1851 at the lower line. of crosses is equal in all cases. The other lines of crosses indicate the noticeably broad rings of 1868 and 1878. An examination of the photographs shows that a very characteristic feature is a group of narrow rings about the years 1879 to 1884. These can be identified in practically every tree and an examination of many stumps which were not measured showed that it was easy to pick them out wherever one chose. Striking verification of this was found in the case of a stump near town which had been cut about 20 years previously. By finding this group of rings, the writer was able to name the year when the tree was felled and the date was verified by the owner of the land. In the more recent work this same group shows conspicuously among Prescott trees, and in general 95 per cent of these trees have rings so characteristically marked that the identification of the same series of rings can be made with little doubt, whether at Flagstaff or at Prescott.

DOUGLASS ⁠ PLATE 2

Cross-identification of rings of growth in yellow pine (Pinus ponderosa).

Fig. 3.—Annual growth of trees at Flagstaff from 1385 to 1906, A. D.

only. A comparison between the 5 and these oldest 2 taken by themselves give an agreement not absolutely perfect, yet so close that errors thus introduced will not materially affect the curves. However, the oldest 2 were very slow-growing trees and they required on the average an increase of about 30 per cent in order to make their curve continuous with the whole 5. Thus the tree-record is made to begin at 1385. In the recent years of the record also, between 1891 and 1896, a slight correction was made for omitted rings, the complete omission of a ring being an exaggeration that introduces error.

Fig. 4.—Comparison of two Flagstaff groups. Variations in annual rainfall according to month of beginning annual means.

THE SECOND FLAGSTAFF GROUP OF SEVEN SECTIONS.

In 1911 the writer visited Flagstaff again and made a trip into the forest where cutting was going on, in order to procure a few additional samples of the yellow pine which would check the recent part of the tree-record previously obtained and bring it up to date for comparison with rainfall values. The location was about 12 miles southeast of town and from 6 to 12 miles east of the region from which the first Flagstaff group was obtained. Seven cuttings were procured from the edges of stumps, thus bringing away a triangular pyramid of wood, which included the outer 50 to 100 rings.

Fig. 5.—Growth of individual trees compared with precipitation at Flagstaff.

Figure 4 shows how well the second group checks the first and indicates that even a small group of trees, no more than 7 in number, is sufficient to give results of considerable accuracy. Indeed, we may go further and say that a single tree under favorable conditions may give results of very great value. This is evident in figure 5, where the 7 sections from the last Flagstaff group are plotted separately, the most rapid grower at the top, just below the rainfall curve, and the slowest-growing tree at the bottom. All rise alike because the conditions of rainfall in 1900–10 were more favorable than in the preceding decade, but all (especially the curve of section 4) show a more or less close relation to the rainfall at Flagstaff, even though that town was some 12 miles away. The great sinuosity which a quick-growing tree may show is well illustrated in section 4 in the great differences between successive years. A lack of sinuosity is shown in section 5 at the bottom. This difference supports the conclusion already reached that slow-growing trees are of less value than rapid ones in the determination of climatic cycles. The results of the measures of this group serve as a check on the preceding measures and are shown in the figures just referred to. They are, therefore, not tabulated in this book.

THE PRESCOTT GROUP.

Prescott is located in the northerly part of the Bradshaw Mountains, at an elevation of 5,200 feet. The rocky subsoil is largely granite disintegrated at the surface and worn into steep hillsides, deep gorges, and picturesque masses of rounded boulders. The ridges are sharp and rugged, and the general contour is very irregular. There are very few isolated peaks. The mountains are covered with pines from their crests to a little below the level of the city.

This Prescott group was obtained in 1911 for the purpose of testing the conclusions derived from the Flagstaff trees some years earlier.

Fig. 6.—Annual growth of trees near Prescott, Arizona.

It consists of 67 trees selected in 5 subgroups depending on their nearness to town. The farthest was 10 miles southeast and the nearest was 1 mile south. It was apparent that the agreement between growth and precipitation increased as the location of the actual rainfall station was approached. The nearest subgroup, containing 10 trees, shows so much greater agreement than the others that it has been used alone in drawing final conclusions. Its site was a small, poorly drained level space near the bottom of the valley.

In this group there was no necessity of duplicating the Flagstaff records, and therefore small V-shaped cuttings were made at the edges of the stumps, only triangular pieces of wood giving the outer half century of ring-growth being brought away. These were the samples on which the value of the cross-identification was discovered, as already described. Identical series of rings were observed in nearly every tree of the group.

Fig. 7.—Annual rainfall and growth of trees (Group V) at Prescott. Dotted line: rainfall. Solid line: growth.

Out of 67 sections averaging 50 rings each, only 6 gave any identification trouble. In 2 of these, 2 rings were lacking, but when allowance was made for this defect the identification was satisfactory. Another section had 2 extra rings, and another had 2 extra and 3 lacking. The other 2 sections proved especially puzzling and were finally omitted from the means. Of these 6 troublesome sections, the first 5 were very slow growers. Hence it would seem advisable not to use extremely slow-growing trees any more than is necessary. It may be urged that trees do not grow continuously at the slow or fast rate and that we can not tell how much of the change is due to rainfall. On the whole, however, it seems advisable to exclude trees or parts of trees whose identification is extremely difficult. The inner rings if well identified may be extremely useful in carrying back early records, as the slow-growing trees are likely to be among the oldest.

The averages of 4 subgroups and the means of all the Prescott trees will be found plotted in figure 6. The curve of the fifth subgroup is given in figure 7, where it may be compared with the rainfall of Prescott.

SOUTH OF ENGLAND GROUP.

This group of 11 sections was obtained in January 1913 at Fleet, near Aldershot, some 30 miles west-southwest of London. The trees were the common pine, Pinus silvestris, and averaged about a foot in diameter. The growth was very rapid and the wood was full of moisture. The trees had formed a border to a little plot of cultivated land with a southwesterly exposure. The average age was 54 years. The rings were all extremely plain, averaging 2 to 4 mm. in size, and cross-identification was everywhere perfect. Of the 50 or 60 rings, about 10 had marked characteristics and were easily recognized in nearly every section. It was noted that a few sections had numerous rings more sharply defined on the summer side of the dense red portion than on the usual winter edge. One of the 11 sections is shown in plate 3, A.

The appendix contains a table of mean tree-growths of the 11 British sections; the years 1859 to 1863 inclusive show means of 6 trees only, as some did not extend back that far; of these, 2 had their centers about 1858, 2 in 1857, and 2 in 1855. The owner of the land informed me that the trees had all been planted at the same time, and therefore this apparent discrepancy may be due to sections cut at different heights above the ground. These means are plotted in figure 8.

For ready comparison it seemed desirable to standardize this British curve as well as each of the other European curves. Each curve is therefore corrected for changing rate of growth with age and also very slightly smoothed to get rid of the confusing effect of the 2-year "seesaw " described later. In the present group, after careful consideration, the standardizing line follows the tree-growth through a uniform curve in the earlier years and becomes straight in the later years. Percentage departures from this mean standard line give the standardized curve. These percentage departures smoothed by Hann's formula will be found plotted in figure 23, together with similar curves from the other European groups.

OUTER COAST OF NORWAY GROUP.

On the advice of Dr. H. H. Jelstrup of Christiania, I visited the Forest School of Sopteland, a small place located about 18 miles south of Bergen, near latitude 60°. The elevation is but little above sea-level, and irregular intervening hills give slight protection from the North Sea storms. This group of 10 Pinus silvestris sections was collected on January 3, 1913, from logs in the yard of the Forest School. The logs had been cut within a week or two in Os, 12 miles to the south, on an exposed part of the coast and probably close to sea-level. Os is on the north shore of one of the larger inlets entering on the north side of Hardanger Fjord.

The average diameter was 6 to 8 inches and the average date of the center was about 1840, but one extended back to about 1800 and another to 1700. The average size of rings was about 1.25 mm. The group cross-identified extremely well and on a preliminary inspection seemed to show somewhat rhythmic variations in growth. In these

	DOUGLASS ⁠ PLATE 3
A
B

A. Section of Scotch pine from southern England.

B. Section of Scotch pine from coast of Norway.

Fig. 8.—The nine European groups.

sections there is more than usual variation in different radii, an excess of growth starting in one direction and then slanting off in some other direction. Here it was found also that maxima were not always the same in different radii. It was suspected that some radii tried to follow a single cycle and others a double cycle. A photograph of one of these sections is shown in plate 4, a.

The appendix presents a table of mean growth of this group from 1845 to 1912. No. 2 had its center in 1865, and between that date and 1845 extrapolated values have been used in forming the means. These extrapolated or artificial values preserve the average shown by the individual tree during its years of growth, but are made to vary from that average in accordance with the variations of the rest of the trees in the group. From 1828 to 1844 the mean of 3 sections only is given. The actual mean has in this latter case been multiplied by 1.25 to bring the average into accord with the group, for the mean of these 3 for the 11 years from 1845 to 1855 inclusive is only 80 per cent of the mean of the group. In this group one center was in 1865, three in 1844, one in 1842, one in 1840, one in 1836, one in 1827, one in 1800, and one in 1693.

These means are plotted in figure 8. The same corrected to a standard mean and smoothed by Hann's formula will be found in figure 23. No real correction for age has been made in this case, for there seems little change in rate of growth that can certainly be identified as such. The whole, therefore, has been simply reduced to scale for comparison with other groups by dividing every year by 1.25, which is very nearly the average growth in millimeters.

INNER COAST OF NORWAY GROUP.

It is a great help to visit the exact locality in which the trees grew, or to get very near it, as in the groups already described, and especially to obtain personal information in a mountainous country like Norway, where meteorological conditions may vary enormously within a few miles. But it was impossible in the present group, whose sections had mostly been collected some years before for use in the forest service and schools. By courtesy of various officials I was permitted to examine and measure these sections in their offices, and whenever it was possible thin sections were cut off for me to add to my collection. In measuring sections of which samples were not retained, for example B 15, B 16, and N 2, there was no opportunity of cross-identifying rings, and hence unusual precautions were observed in numbering the rings. If at any spot they seemed to be very close together with any chance whatever of mistake by omission or doubling, the numbering was carried to as many other radii as were necessary for a check, and worked over very carefully until the best possible result was obtained and all doubt seemed to be overcome. Nevertheless, judging by past experience, unchecked counting leaves a doubt wherever the rings are reduced to 0.1 to 0.2 mm. in thickness.

Another disadvantage of this group is that the trees came from very diverse localities, and hence do not represent homogeneous conditions. Therefore, each section in the group will have special mention. The first number in the group, B 11, was cut from a log of Pinus silvestris lying on the woodpile in the yard of the forest school at Sopteland. The tree had been brought in for firewood late in 1912, but was undoubtedly dead at that time, for the outermost ring checked with the Ös group unmistakably as 1911. This view was supported by the decayed bark and moldy trunk. This section was 9 by 14 inches in size and had the center (date 1734) some 3 inches from one end, producing one of the most uniform cases of eccentric growth which I have seen.

Nos. B 12, 13, and 14 cross-identify most satisfactorily with the Ös group. From one to six individual characters or a most convincing sequence of characters were obvious in every decade. Section 12, a foot across, was cut in 1909, and the last complete ring was unmistakably of 1908 by comparison with the previous group. A section was cut for me at the school in Sopteland. The original was marked "No. 1, 1909, Knagenkjelm, Kaupanger," a location on Sogne Fjord, some 80 miles northeast of Bergen. My section shows the bark and very dense, handsome wood with strongly marked rings. Its center is at 1682. No. B 13, center at 1807, is of about the same size and from the same place, and was marked "No. III, 1909." Its outer ring also identified as 1908. A portion of this section also was cut for me.

No. B 14, center 1779, was marked "No. 1, 1909, Lyster Sanatorium," on Sogne Fjord. As in the other two cases, its outer ring was plainly 1908. A thin section was cut for me. Its size was 12 by 14 inches. B 16 was marked "No. 3, 1909," from the same place. This huge section was 28 inches in diameter and 7 inches thick, and its center was about 1724. There was a series of very small rings from 1787 to 1794 and another from 1806 to 1813. I have no section of it and so no cross-identification could be attempted, but the measures of the recent years agree with No. 14 from the same place.

No. B 15, center at 1633, was also measured at the school and no section retained for comparison with the others. It is the only one from its locality. It was marked "No. 1, 1909, Nestaas, Granvin," on Hardanger Fjord. It was cut in October and the first ring was considered to be of that same year. The rings were very clear back to 1680 and in fact to the center, but between the center and 1680 they were very small. All the sections so far in this group came from the west side of Norway near latitude 60°. The remaining two came from farther north and were first examined in the office of Dr. Jelstrup. No. N 1 was a small tree some 6 inches in diameter with its center in 1848. It grew in Mo i Ranen in latitude 66° 15', a 2 days' trip by boat from Trondjem. The rings show a rhythmic character, and a photograph of the thin section presented to me is given in plate 3, b. As in the other similar photographs, the years of sunspot maxima are marked with arrows. It was cut in 1907 and the outer incomplete ring was taken as of that year. The identification with trees from near Bergen is poor, as would be expected.

No. N 2 was an interesting cross-shaped section from beyond the Arctic Circle, latitude 68° 45'. It had been damaged by forest fires at various times to such an extent that the injured parts of the trunk ceased growing while the rest kept on; hence it was of this extraordinary shape. It was cut in the winter of 1905-6, and the outer ring was taken as of 1905. As a rule the rings were very easy to follow until before the year 1600, and even then by carrying the ring to other arms the identification seemed practically certain. The rings reached a suspiciously small size between the center at 1497 and 1512.

The measuring of this 400-year section was done on December 31, 1912. By noting ring after ring with care, tracing all rings a short distance and following the one case of suspected double across into another arm, there seemed to be no errors, certainly none of doubling and none suspected of disappearance. Letters B, BB, indicating maximum growth, were placed at the center of groups of large rings as the measuring progressed, without knowledge of any relation between them. That same day, on looking over the measures, a Brückner period seemed indicated. The maxima were marked as the measuring progressed. See table 3, on page 35.

This series of maxima, 270 years long, from 1561 to 1830, shown in figure 38, permits the application of a 34-year period with an average error of less than 3 years. If that case were alone, I would not include it here, but I believe I shall be able to show it in a number of very old trees in widely separated localities.

From the above description it is evident that we have in this group some very interesting trees, even though they grew far apart. They are probably worth more as individuals than as a group, but until more trees can be added from their various localities the usual method of presenting them here is used. So the group means are tabulated in the appendix, using an extrapolated value of N 1 from its center in 1848 back to 1821. These means will be found plotted in figure 8. They have been corrected for age and reduced to standard size in the usual way by a straight sloping line reading 1.90 mm. in 1820 and 1.15 mm. in 1910. The corrected means smoothed by Hann's formula will be found plotted in figure 24.

Table 3.

Date of maxima as marked.	Differences in years.		Suggested maxima on 34-year period.		Residuals.
1830		..........	1831		-1
1797		33	1797		0
1754		43	1763		9
1696	$\scriptstyle {\left\{{\begin{matrix}\ \end{matrix}}\right.}$	29	1729	$\scriptstyle {\left.{\begin{matrix}\ \end{matrix}}\right\}\,}$	1
		29	1695
1657		39	1661		-4
1528		29	1627		1
1561	$\scriptstyle {\left\{{\begin{matrix}\ \end{matrix}}\right.}$	33	1593	$\scriptstyle {\left.{\begin{matrix}\ \end{matrix}}\right\}\,}$	2
		34	1559
1535		26?	1525		10?

 

CHRISTIANIA GROUP.

This group of 5 Pinus silvestris sections was secured from logs at a little sawmill in the outskirts of Christiania. The logs cut in the neighborhood were in a large pile at the mill, and after the snow was brushed from them suitable ones were selected. Usually in such cases the largest and oldest were taken, but in this group the growth was exceptionally complacent. Accordingly, preference was given to those which showed variability in size of rings.

These sections were measured a month or two later. Cross-identification proved very unsatisfactory. Large variations were found in the 5 specimens. On this account it was felt that there might be several errors in this group which could perhaps have been removed by a larger number of trees for intercomparison. The centers of the 5 were respectively at 1848, 1824, 1797, 1807, and 1790. The average diameter was about 1 foot. On page 114 will be found the mean growth of these sections and the plot of the same will be found in figure 8. There seems no special change in growth with age, and the whole series was merely reduced to percentages by dividing each yearly value by 1.50 mm. These values, smoothed by Hann's formula, will be found plotted in figure 23.

CENTRAL SWEDEN GROUP.

These 12 sections, showing an average diameter of about 11 inches and an average age of 190 years, were obtained from the sawmill near Gefle, on the coast, 60 miles north of Stockholm. The mill, one of the largest in Sweden, was some 4 miles from the town, on the river coming from the interior. The logs came from the vicinity of Dalarne in central Sweden, a large district. The bark is taken off as required by law and the logs are floated down to the mill. I visited the mill on Saturday, December 28, 1912. Twelve sections had been cut, but they were too thick and the whole 12 were cut a second time. These logs had been in the water a year and the last ring would therefore be of 1911 or possibly 1910. Of the 12, I think that all but 2 or 3 show the 1911 ring. Though these sections must have come from a considerable area (unless in the water and mid-afternoon darkness they accidentally secured original neighbors from thousands of logs), they identify among themselves extremely well. Cycles or pulsations were noticed and marked on all the sections of this group before identification. No. S 8 seems the most regular; a photograph of it is reproduced in plate 4, b. The cross-identification for the last 100 years hardly needs review, as it is entirely reliable and practically nowhere are there doubtful rings.

The means of the years 1820 to 1910 are given in the appendix and a plot of the same will be found in figure 8. The tree-growth in this group and others before 1820 will be taken up separately. There seems to be here no real change of growth with age, and the values were changed to standard by dividing by 0.8 mm. These results were then smoothed by Hann's formula and plotted in figure 23.

SOUTH SWEDEN GROUP.

This group of 6 sections was measured at Stockholm on December 27, 1912, in the office of Professor Gunnar Schotte, chief of the Swedish Forest Service. In my lists they are numbered from S 13 upward. They are all Pinus silvestris save S 14 and S 17, which are spruce, Picea excelsa. No. S 14 is noted particularly because it showed as perfect a sun-spot rhythm as G 8 from Eberswalde, whose photograph is given in plate 8, a. An entirely satisfactory cross-identification was made at the time of measurement.

The individual trees came from different localities and are therefore mentioned separately. No. S 13 was marked "4105-6" and was cut in May 1909. It grew about 100 miles southwest of Stockholm, in latitude 58° 40'. About 1833 it has a doubtful ring which was settled by comparison with other measures. Its center was in 1763. No. S 14, a Picea excelsa, marked "4105-14," was cut in July 1910 on the east side of Vetter Lake, less than 100 miles southwest of the preceding. Its center was in 1816. No. S 15, marked "4105-2," was cut in August 1909, about latitude 64° 30', near Lycksele, Lapland. It showed clear and well-sized rings to its center in 1701. No. S 16, marked "4131-a1," was cut in August 1910, in Elfdals, in latitude 61° 24'. Its center was about 1838, but its inner 10 rings were uncertain and therefore not

	DOUGLASS ⁠ PLATE 4
A
B

A. Section of Scotch pine from Ös, Norway.

B. Section of Scotch pine from Dalarne, Sweden.

used. No. S 17, Picea excelsa, marked "4105-5," was cut in May 1909, in the same locality as S 14 and shows a similar rhythm. Its center was in 1777. No. S 18 was a small section marked "4131-a-12." It was cut in October 1910, in latitude 58°, well to the west of the others. Its radius measured only 2 inches. The average diameter of the other sections was about 12 inches.

On page 115 will be found the means of these sections, with two extrapolations, one from 1820 to 1848 and the other from 1820 to 1878. This curve will be found plotted in figure 8. It has been corrected for age and reduced to percentages by dividing by the readings of a straight line extending from 1.90 mm. in 1820 to 0.70 mm. in 1910. This corrected set has been smoothed by Hann's formula and will be found plotted in figure 23.

EBERSWALDE (PRUSSIA) GROUP.

These 13 trees were cut and sections prepared for me by the kindness of Professor A. Schwappach of Eberswalde. They were all Pinus silvestris planted about 1820 to 1830, exactly alike in height and size, with tall, straight, clear trunks about 10 inches in diameter and bushy tops. The land is a gently rolling country with a slight northerly slope, leaf-covered ground, a sandy soil with loam on top, and an elevation above the sea of 200 to 300 meters. The height above the city level was 200 feet or so; the locality was south and west of Eberswalde station. The trees cut were scattered along a quarter of a mile and so did not represent any close grouping. Their rings show almost identical records; 2 to 10 in every decade have enough individuality to make them recognizable in every tree.

On the first examination of these sections in November 1912, it was evident that their growth follows with fidelity the sunspot curve since 1830. This may be traced in the curves below and in the accompanying photographs of two of the sections in plate 8. It will be seen at once that there is a rhythmic sway in the growth, groups of large rings alternating with small ones. The arrows placed in the photographs mark the years of maximum sunspots. Taking the group as a whole, the maximum growth comes within 0.6 year of the sunspot maximum. To one maximum alone they fail to respond, namely, 1894; instead of rising, the curve drops in 1892, 1893, and 1894. I have tried to find cause for this, but was informed by Professor Schwappach that there were no fires, pests, or other known causes for it except climatic conditions.^[2] On page 114 will be found the Eberswalde means from 1830 to 1912. Only one center occurs later than that date of beginning, namely, 1833. The others were mostly between 1821 and 1827. The means of this group are plotted in figure 8 and also shown more in detail in figure 9. These means have been corrected for age and reduced to

Fig. 9.—Sunspots and growth of trees at Eberswalde, Germany.

percentages of a mean line reading 2.57 mm. in 1830 and 0.54 mm. in 1910. These in turn have been smoothed and plotted in figure 23.

In considering the significance of the agreement above noted, one should, in my opinion, keep in mind first the unusually homogeneous environment of these particular trees and the great care they have received, and second, the suggestion they contain of eventually defining distinct meteorological districts in which homogeneous effects are noted. A small pine of 60 rings from the Hartz Mountains was examined in the Geological Museum at Berlin, in which the same cycle was prominent. It was not measured, as the date of cutting was not known.

PILSEN (AUSTRIA) GROUP.

This group of 7 Pinus silvestris sections was measured in the office of Dr. A. Cieslar, in the Hochschule der Bodenkultur at Vienna. I have no samples of them in my collection, but they were carefully cross-identified before measuring. Two sections had the ring for 1849 very doubtful, but its identity was verified by comparison with the others. The average date of the center was 1821 and the average size 11 inches in diameter. They all came from a forest station near Pilsen, in northwestern Austria. I have not seen the locality, but judging by the appearance of the country a little farther south the mountains are not rugged.

The mean measures upon this group from 1830 to 1912 are given on page 114; their plot is in figure 8. The curve is very peculiar, and it is hard to say how it should be corrected for age. It descends sharply from 3.62 mm. in 1830 to 1.25 mm. in 1851, and from that point on it remains 1.25. This bent line has been applied, and the resulting percentages have been smoothed and plotted in figure 24.

SOUTHERN BAVARIA GROUP.

This group of 7 Pinus silvestris and 1 Picea excelsa sections was cut for me by the kindness of Messrs. Klopfer and Konigen in Munich from logs in their yards. The trees had been cut in the winter of 1911-12 at Altotlinz, Ober Bayern, some 50 miles south, at a considerable altitude, in the northern valleys of the Alps. The rings in all these were clear and distinct and no doubtful cases were found except a very few near the center of two sections, which were omitted in the means; yet the cross-identification was not fully satisfactory. Sections numbered M 2, M 6, and M 8 in this group showed 1 to 2 entire discordances out of about 10 features in the last 60 years. The others agree fairly well. Possibly this condition results from the rugged and non-homogeneous region where they grew.

On page 116 will be found the means from 1848 to 1911. In these, M 6 and M 7 are extrapolated for about 12 years, and M 8 for 2. These means will be found plotted in figure 8. The curve shows apparently a very rapid decrease of growth with age. The correction line assumed is a line reading about 3.15 mm. in 1850, 2.10 mm. in 1860, then with decreasing slope reaching a nearly level line at 0.90 mm. between 1895 and 1911. The means have been reduced to percentages of this line and smoothed by Hann's formula and plotted in figure 24.

OLD EUROPEAN TREES.

It is of course most desirable to carry the tree-records back as far as possible for verification of any feature observed in recent years and for additional information. But one is met by the rapidly diminishing number of specimens and the liability of obtaining records which are not representative of the regions on account of the increasing effect of individual and accidental variations. It is true that in the very homogeneous region about Flagstaff, Arizona, an average of 5 trees and even of 2 gave a valuable record corroborated by comparisons with larger numbers; but in these European groups the oldest trees are all from the Scandinavian peninsula, and probably the individual trees of which I have samples are representative of widely different localities in a rugged and mountainous country. Even though not homogeneous, the 15 oldest trees have been segregated in 2 groups covering the interval from 1740 to 1835.

Group A represents the inner coast of Norway and includes the following trees: No. B 3, Ös, south of Bergen; No. B 11, Sopteland, south of Bergen; No. B 12, Sogne Fjord; No. B 15, Hardanger Fjord; No. B 16, Sogne Fjord; No. N 2, latitude 68° 45'.

Fig. 10.—Growth of old European trees. A, six Norwegian trees, mostly from inner fjords. B, eight trees from Dalarne, Sweden.

Group B is made up of 8 trees from Dalarne, central Sweden, and 1 from Lapland, latitude 64° 30'. This group, therefore, represents somewhat more homogeneous conditions, but yet it can not be well summarized in its larger fluctuations. When plotted with Group A, as in figure 10, it shows the latter to have a considerable tendency to reversal, a characteristic already observed in this region. But there are discrepancies in Group B consisting of sudden depressions in growth which suggest injury, as in 1756 and 1769 and 1770. A very regular recovery from these depressions sustains this idea of their cause. The means of the 2 groups are given on page 116.

A few trees are perhaps available for periods antedating 1740. The centers of 5 are as follows: B 12 in 1682, B 15 in 1641, N 2 in 1497, S 4 in 1510, S 9 in 1660. But the first 3 are from separate localities in Norway and the other 2 are from central Sweden, so it seems hardly profitable to include them here in a group on account of the tendency to reversal between those localities. The section N 2, 400 years old, from high latitude on the Norwegian coast, presents a feature of interest as noted in connection with the Norwegian group, namely, a pronounced fluctuation very nearly 34 years in length. The measures on this tree have been plotted, a mean sinuous line drawn through them, and then this mean line transferred to a different scale, smoothed graphically, and photographed to form figure 38 on page 106. The more formal analysis of this interesting tree-record with the periodograph confirms this periodic fluctuation.

WINDSOR (VERMONT) GROUP.

On return from Europe it seemed desirable to learn how American trees react in similarly moist climates. But it was not easy to secure sections. There are very few large pines near the Eastern cities. One "pitch" pine from 50 miles south of Boston, with more than 100 rings, was secured, but there were no others in that immediate vicinity. Five white-pine sections from near Middleboro, Massachusetts, were obtained, but their rings were too few in number, being only 50 to 60. Finally a satisfactory series of hemlock, Tsuga canadensis, from Windsor, was collected. Six sections came from the northwest slopes of Mount Ascutney at the lower and very steep end of the Brownsville trail. Five of these I cut from the stumps myself and preserved, and one was measured on the stump itself with full cross-identification. The remaining 5 of the 11 were cut from logs in a lumber-yard in Windsor; they came from across the river on a farm about 3 miles from town. Thus 7 or 8 miles separated these two subgroups. But the whole are here retained in one group, for the cross-identification, though difficult, was perfectly satisfactory. In order to be quite sure on this point, the subgroups were left separate until their curves could be compared. The Ascutney subgroup, with one extrapolation, extends back to 1695, and from that date 2 trees were carried back to 1650. A comparison between the 2 and the whole 6 showed harmonious curves in their overlapping parts. This curve shows an average growth of considerably less than 1 mm. in all its earlier years and up to the year 1808, when its yearly growth doubled. This sudden increase was interpreted to mean that at that time these hemlocks emerged from the shade of surrounding trees. The change was so rapid and great that it seemed likely to be due to the cutting down of the surrounding forest. In this subgroup, also, the years 1770 and 1821 were so extremely small that injury on those dates seemed likely.^[3] The other subgroup from east of Windsor extends easily to 1650, with one extrapolation of 20 years and another of 3. It shows no effects in 1770 or 1821, but does show a temporary slight rise in 1807, and then a gradual increase to well over 2 mm. by 1870 or 1880, as would be expected when light-loving trees gradually push their way out into preeminence above their neighbors. A comparison between these two curves in their minor details confirms the view that all 11 may be included in one group.

The means of the Windsor hemlock sections from 1651 to 1912 are given on page 116. In 1651 the figures give an average derived from only 6 sections. This increases to 9 sections in 1694, and from 1695 the whole 11 sections are used. These numbers have been smoothed and plotted,. and their resulting curves will be found in figure 27, together with the sunspot curve.

OREGON GROUP.

Following the New England group, a set of Douglas firs was obtained from a logging area about 25 miles northwest of Portland, Oregon. Several points of interest appear in connection with this group. In the first place, the samples were not radial specimens of the wood itself as heretofore, but were pieces of blotting-paper of suitable size which had been rubbed into the tops of the weathered tree stumps. These were made in 1912 by Mr. Robert H. Weinknecht, who writes as follows:

"The prevailing age on the tract is about 210 years on the stump. The trees selected were average with neither suppressed nor abnormally large growth. An average typical radius was selected on each stump. Twenty-three impressions from this one locality were obtained and sent. Twenty-one came from stumps cut in the summer of 1908, one from a stump cut in 1909, and one from a stump cut in 1912. The method of taking the impressions was one devised by Mr. Higgs and described by him in the Forest Quarterly for March, 1912. It was found that fresh stumps gave very poor results and especially poor for the last 50 years. This was attributed to pitch forming near the outer parts of the stump and to the fact that the weathering of the stump had not been sufficient to bring the rings out in relief. Some of the impressions were gone over with a pencil to bring out the rings where they were faint or broken. This was done carefully and checked by the number of rings counted on the stump."

During the course of identification and measurement, it was observed that only a small proportion of the ring impressions show distinctly the ending of each year's growth, which is usually the best measuring-point. In good reproductions the measures are satisfactory, different observers agreeing within 0.1 to 0.2 mm. In others, however, there is much chance for judgment in selecting the measuring-point, and observers differ 0.3 to 0.5 mm. Nevertheless it is easy to judge of the relative sizes of rings and the only injurious effect is to reduce variations.

The cross-identification was very satisfactory, with practically no doubtful cases and only a few which required careful study. It is not likely that there is a single error in identity throughout the 17 sections in this group. Two other trees, one cut in 1909 and the other of

Fig. 11.—Oregon group. Curve No. 1, actual tree-growth; No. 2, tree-growth departures, smoothed; No. 3, sunspot numbers displaced 2 years to left.

unknown date, but probably cut in 1902, were not included. They showed special characteristics, such as an evident injury in 1861-62, affecting the 1862 ring and several others following it. They show also small growth in 1886, and even in 1887 and 1888 following the minimum growth of 1884 and 1885 prominent in the large group. The tree (cut probably in 1902) shows a minute growth in the years 1779 to 1783 inclusive, evidently the result of injury. These two sections are full of character and may prove valuable.

Five other rubbings similar to the group of 17 were discarded because defective in some parts. The attempt to trace the lost lines with a pencil-mark gave no help. One of the 17 was defective since 1835, and only the earlier part, ending in 1834, was used. Extrapolated values for the missing part were derived in the usual way. A few short, apparently doubtful, regions of rings required careful study and it was found that well-adjusted illumination of the rubbings was very necessary to their correct reading. When the ring impressions were deep in the paper, the end of the rubbing showing the tree center was held toward the source of light in order that the elevation corresponding to the beginning of the spring growth might be brightly illuminated. When the impressions were shallow and faint, it was noted that the rings became very distinct if the rubbing was held between the eyes and the light, thus giving a very faint and perfectly even illumination. If this did not bring out the individual rings, the rubbing was not used.

The location in which these trees grew was visited in 1918 and general contours were noted. The hills are low and comparatively flat-topped, with disintegrated rocks showing in railroad cuttings. The sides of the hills are steep, and the valley bottom is narrow and usually has a wash near its center. In general the drainage is toward the east, but there is no high and sharp ridge between this region and the ocean on the west. The situation is far enough north to have a good snowfall in winter. It is about 800 feet above sea-level.

The tabular matter giving the results of the measures on the 17 Douglas firs of Oregon will be found on page 117. The plotted values appear in figure 11.

THE SEQUOIA GROUP.

In 1911, after examining the writer's results obtained on the yellow pines, Huntington made an extensive series of measurements on the big tree, Sequoia gigantea. He did this work on the stumps themselves by direct counting from the outside. This introduced errors of beginning due to removal or injury of outer rings, and errors of omission which of course could not be checked. In order to correct for large errors of omission, he worked out an approximate correction on the grounds of probability which depended upon a comparison between two or more radii of the tree, and in that way many errors were compensated. In the vast majority of cases, his measures were not of individual rings but of successive groups of ten. I have collected seven of his trees, and after complete cross-identification verify his centers as shown in table 4.

But Huntington's method of working directly on the stump enabled him to get data from a very large number of trees, some 450, in a way that served his purpose very admirably. He was searching for general effects, and accuracy to a year or two was less essential. He wished to approximate absolute values of rainfall in past climates, in contrast with which my chief aim is to get relative and periodic values. These two different purposes supplement each other in a highly valuable manner. Therefore, for him, the determination of the general curve, with an allowance for larger growth near the center, was most important. For that purpose he used both young and old trees. Necessarily he visited places where the trees had been cut. The two chief regions of his measurement were in the King's River Canyon district close to the General Grant National Park, and in an old lumber region near Springville, which is south of the Sequoia National Park.

Following Huntington's route, I visited the former region in August 1915. The town of Hume, the mill-site of the Sanger Lumber Company, is reached from Sanger by daily auto stage and formed, therefore, an excellent base of operations. Hume is at an elevation of about 5,500 feet, on the shore of a large artificial pond, into which the logs are dumped as they are brought down from the camps. A narrow-

Table 4.

Sequoia No.	Huntington's No.	Huntington's first year of tree.	Identified first ring.	Distance from center in inches.	Probable date of center.
12	92	17 A. D.	Not ident.	8(?)	...
13	91	585 A. D.	588 A. D.	0	...
14	96	387 A. D.	389 A. D.	0	...
15	59	121 B. C.	159 B. C.	0	...
21	74	1318 B. C.	1304 B. C.	1	1316 B.C.
22	195	1141 B. C.	1086 B. C.	7	1160 B. C.
23	116	1191 B. C.	1121 B. C.	10	1200 B. C.

gage logging road extends in an easterly direction from Hume, high up on the southern side of King's River Canyon. It winds in and out of the various small canyons or basins that empty into the large ravine. The elevation of the log road increases gradually from Hume until it reaches 7,000 feet at Camp 6 and Camp 7, which are about 7 and 9 miles distant respectively.

Camp 6 and Camp 7 are the names of the two recent logging stations. Camp 6 was occupied in 1915 and was located on the eastern side of Redwood Basin. The camp sites are usually chosen in such localities, for in each basin there is an enormous collection of accessible timber. In general the tops of the mountains are very rugged and the slopes exceedingly steep. The upper ridges are apt to be very sharp, but in the higher altitudes there is a tendency for the weathering of the mountain to produce this basin type of contour. From the accumulation of soil and the enormous snowfall in winter these become exceedingly swampy. Below the basin the water is carried by sharp, narrow canyons down very steep grades to the river far below. These groves of sequoias are between 6,000 and 7,000 feet above the sea. The climate at this elevation presents a contrast between an intensely cold winter season with 10 to 15 feet of snow and delightfully mild summers. The latter have occasional thunder-storms whose waters quickly run down the mountain slope. Thus conservation plays an important part in the growth of these trees by rendering the winter precipitation more important than the summer and by permitting the moisture to remain long in the swampy places.

Three groups were obtained from this general region in 1915. The first of these came from the uplands above Camp 6 close to the west line of section 17, township 13 south, range 29 east. This region may be found on the Tehipite Quadrangle of the United States Geological Survey. The group includes Nos. 1 to 5. No. 1 was a splendid tree, about 19 feet in its greatest diameter, growing at the uppermost limit of the logging area. Its growth was rapid, and yet it was an extremely sensitive tree, showing beautiful variations from year to year. No. 2 was obtained a little lower down and is mentioned here because it has been used as the standard of the whole sequoia group, having probably a more perfect record than any other tree measured. Its center was about 300 B. C. No. 5 was a small tree which was cut just at the time I came within hearing distance. I thought that two blasts of dynamite were set off and found afterwards that only one charge of dynamite had been used to break through the last support of the mighty tree; the other report was the tree itself crashing to the ground. Yet this was a small tree, only some 12 feet in diameter, and its age was about 700 years. It proved of particular value to the whole sequoia group, because it was the only tree on which was obtained the ring of the current year, thus permitting a very important correction to be made in the dating of rings. This had an important bearing on the relationship of rings to rainfall.

The second group included Nos. 6 to 11, and was made about a mile to the north and 700 feet lower altitude in the swampy basin whose outlet was similarly toward the northeast. No. 6 grew at the edge of the little brook running through the basin and its rings proved later very uncertain in identity, because its habit was complacent, i. e., the rings were nearly all alike in size.^[4] No. 7 was an improvement on it, and No. 8, which was still farther from the creek, was perhaps the best of this group of 6. It gave a very fine cross-identification with the first group. No. 11 was also very close to the creek near the outlet of the basin and, as with No. 6, it was impossible to be sure of the identification, owing to its complacent character.

The third group consisted of 4 trees from Indian Basin, about 10 miles northwest of the Redwood Basin and 3 miles north of Hume. This basin is a broad, flat, fertile area with an outlet toward the northeast. Four trees were obtained there which Huntington had already counted. Nos. 12 and 13 came from the flat middle area of the basin. No. 12 was not included in the final averaging because its rings

	DOUGLASS ⁠ PLATE 5
A

B

A. Upland contours, above Camp 6 in Sequoia Grove: D-19.

B. Basin contours, Indian Basin, looking S. E.: D-12 and D-13 in center.

could not be identified at all, chiefly owing to large numbers of compressed rings in the last 500 years or more, and to several heavy fire-scars and its generally complacent character. In 1919 a short radial sample was cut from another part of the stump and a complete and satisfactory identification obtained. It shows very fine rhythmic growth in places. No. 13 was not included in the final averages, because its rings were very complacent and perfect identification was not obtained. Nos. 14 and 15 were obtained from the northern side of the valley and their identification was entirely satisfactory. The agreement which they give with Huntington's "first year of tree" has already been quoted.

The three groups whose collection has been described above showed on examination certain interesting relationships to the location in which they were found. The first group was obtained high up on a hillside, where the slope of the ground was 15° to 25°. It was not very far from the top of a sharp ridge and there was no opportunity for moisture to collect and remain for long periods on the soil. Therefore one would expect these trees to show variation related to the amount of snowfall each winter, if any did. The growth of some of these trees was large but full of constant variation, and they were therefore of the type which I have called "sensitive." They do in fact show best of any the relationship to precipitation which will be described in a later chapter. The second group came from a characteristic feature of the country, namely, a basin with thoroughly water-soaked soil.

The luxuriance of vegetation in these basins before lumbering was wonderful. The sequoias grew often within a few feet of each other, and even between them were pines, firs, and cedars. Lumbermen often point out the bottom of a basin and say that such a place ran over 1,000,000 board feet to the acre. To-day nearly all the trees are gone and debris and rubbish are scattered about everywhere. The constant supply of water in the basin made the trees less dependent upon the annual precipitation and they show, in fact, large rings with very slight variation from year to year. They are typical examples of the "complacent" habit. Complacent trees contribute much less to a knowledge of climatic variations, and some of them have to be discarded because of uncertainty in the dating of their rings.

The third group, Nos. 12 to 15, came from Indian Basin, where logging had been done about 1903. Its outlet, like the others, was toward the northeast. It had, however, a much larger flat area, now covered by extensive fields of hay and by forage. The characteristics of the trees found here were the same as in the groups already described.

No. 1 (with a 7-foot radius) was first counted and marked with provisional dates. The rings were coarse and the numbering seemed promising, but proved later to have 6 to 8 errors in the last 700 years. No, 5, which was the tree cut down during my visit, was then dated provisionally. It was 700 years old, with coarse, sensitive rings, and was the only one of the group showing the ring for 1915. In comparing these two for larger variations no accordance was recognized and in details cross-identification failed also, due (as afterwards found) to accumulated errors in No. 1.

No. 2 was then counted and compared with No. 5 with apparent certainty and satisfaction. The former was nearly 6 feet in radius, with small rings, 2,200 years old, and with all but 3 years represented. The last 700 years were thus compared minutely with No. 5 and the earlier parts with No. 1, and one ring (later identified as 699 A. D.) was found to have been overlooked. The earlier parts were later all checked

Fig. 12.—Cross-identification in first five sequoias and gross rings in No. 1.

against No. 3 and no suspicion of error was discovered. This number was, therefore, taken as the best type of specimen of this group. Large fluctuations of size rarely occur in it.

No. 3 was next counted by comparison with No. 2. No. 3 has few large fluctuations and large portions of it match No. 2 with the greatest accuracy. Nevertheless, as a standard with which to compare others, it would be misleading, for it frequently omitted rings; in one place 7 rings and in another 6 rings are entirely missing, and half a dozen more in singles and in pairs. Yet cross-identification with No. 2 was easy and perfectly convincing as to the location of the missing rings.

No. 4 proved to have fairly large rings with 3 to 4 single ones missing and some hard to find, but the identification was easy and entirely satisfactory. No. 1, which was by this time recognized as the most difficult of the group, was reexamined in detail by comparison with No. 5, which proved difficult, with No. 2, which was somewhat better, but especially with No. 4, which proved to have the closest similarity, and all apparent errors were removed. It was very apt to drop out completely rings which were a little below the average. No. 5 seemed to have no tendency to subdue or drop rings. This, with its disclosures of the ring for 1915, showed the necessity of including younger trees in any new group to avoid mistakes in the outer slow-growing parts of the older trees. A comparison of the last 70 years' growth of sections 1 to 5 is given in figure 12. An illustration of "gross" rings is seen in the upper curve.

When the second subgroup was compared with the first, two complete omissions from No. 2 and the others of that first subgroup were discovered. This necessitated the complete renumbering of the first five sections.

The sections were measured at this stage of the dating process. The final renumbering was made after the 1919 trip, the purpose of which was settling the identity of a doubtful ring occasionally found between 1580 and 1581. The existence of this ring was established and the necessary corrections on the sections and in the tabular matter in this book have been made. All subsequent comparisons have verified this identification.

THE SEQUOIA JOURNEY OF 1918.

The visit to the Big Trees in 1918 was for the purpose of procuring material so that the tree-record from the 2,200 years already secured could be extended to 3,000 years. It was expected to do this without great difficulty, for Huntington had enumerated 3 trees over 3,000 years of age, and he had placed numbers on the tops of stumps so that these could be readily identified. Nevertheless, in consequence of the occasional absence of a number on the top of a large stump which had been counted by him, a little more care proved to be necessary than was anticipated.

After procuring an outfit in San Francisco, I selected Hume as a base and immediately went out on the log road to Camp 6, the old location of the groups obtained in 1915. All the stumps from which samples had been taken (including Nos. 1 to 15) were visited and each was marked with its respective number preceded by the letter D. This marking was done by a chisel, and the figures were usually about 4 inches in height. Placing the capital D before each number made it certain that no number would be accidentally read upside-down. Naturally the stumps from which samples have been taken show the large cut from center to outside, and there is no doubt about their belonging to the group. But if other samples are taken in future years, this numbering will prevent confusion. All the 23 stumps are thus identified by a number in this series.

I had hoped on this trip to find other trees as old as Huntington's three, and therefore searched carefully for the largest stumps. All those over 20 feet in diameter and a number of less size were estimated for age. This was done by measuring the average width of rings here and there along a radius and multiplying by the length of the radius. About 50 were thus tested. In many cases the result has proved to be within 50 years and sometimes much closer, but these estimations were not very reliable, there being several large mistakes in them. In attempting to pick out the oldest stumps among several thousand without spending much time or getting very far from camp, it is impossible to make these estimates with very great care. It was felt that much help would have been obtained from a small range-finder and telescope, the former to give the distance of the stump and the latter its diameter. In the course of a few days this would have saved many miles of tramping and the oldest trees would have been found more readily.

On the steep upland slopes above Camp 6, two trees were estimated at about 2,500 years in age. These were afterwards numbered D 18 and D 19. D 18 was an immense tree which was cut down in 1914 at the time a motion-picture company was operating in the sequoia forest. It is referred to by the lumbermen as the "Moving Picture Tree." It had to be blasted from the stump before it fell, and the stump was so completely shattered that no sample could be cut from it. In falling, the trunk of the tree split in halves through a large part of its length, and most of it remains where it fell. About 40 feet of logs were cut away between the ruins of the stump and the rest of the tree. Accordingly my sample was cut from the lower end of the broken top and at a point which had been about 50 feet above the ground.

Close by the location of No. 18, and on the steep upper hillside just below the track which extends on to Camp 7, is No. 19. A log from it rests uphill with its upper end at the railroad embankment. The section was taken from the stump nearly 60 feet below (see plate 5). Camp 7 was visited and used as a base for two days. It is 2 miles beyond Camp 6 on the ridge at the farther side of Windy Gulch. There are some very fine stumps close to the road that goes down from the camp into the basin, which were estimated to be 2,300 or 2,400 years old.

No. 16 was found high up in the gulch that extends toward the top of the mountain just south of the camp. The gulch faces toward the east and at the location of the tree has a slope of 15°. No. 17 comes from the basin some hundreds of feet below the camp. It was a wind-fall and the lumberman thought it might have been lying there a great many years. As it was a very large tree and of slow growth, it was

	DOUGLASS ⁠ PLATE 6
A
B

A. Cutting radial sample from end of log, Converse Hoist: D-20, age 2800 years.

B. Site of oldest tree, Converse Hoist: D-21, age 3200 years.

hoped that its center would prove of very great age. But the results were disappointing, for it turned out that it had fallen only a few years before the logging began and that its age was only 2,200 years. It had so many compressed rings in its outer parts that the last 800 years were not considered worth measuring.

On leaving the vicinity of Hume several days were spent at the General Grant National Park. It formed an ideal center for a considerable region. Horseback trips were made to the area which Huntington calls the "World's Fair District," "Converse Hoist," and by other names. No. 20 was a fallen tree with a northerly exposure, on the west side of the upper basin, not far from the old hoist at the top of the ridge. It was on the west side of the abandoned railroad. It was found that the tree fell only 6 years before the logging was done. A log had been taken out and the sample was cut from the top of the fallen stump. No. 21 is the most interesting of all, because it gives the oldest record by nearly 200 years. It is on the east side of the railroad and brook in the lower part of the upper basin, and some 30 feet above the level of the brook. It is not at all impossible that during its long life the topographic character of the ground about it has altered materially. It is somewhat complacent in its later growth, but this does not persist throughout its record. The top of the stump had carbonized, become extremely brittle and very hard to cut. Though bits of wood broke off and clogged the saw, every piece was marked and preserved. The radial sample has been glued together in the laboratory and is now 9 feet long. The original center of the stump was badly cracked through contraction in drying, but there were lacking only about 2 inches at the center. The central portion, perhaps a foot in diameter, was not firm enough to be cut out with the saw. It was therefore removed very carefully and is now mounted in a special box in the laboratory. The oldest complete ring in good condition was identified as 1305 B. C. Possibly two more rings may be added. A hundred yards to the south and slightly higher up the hillside is the "World's Fair Stump." This was cut in 1892 at a height of more than 20 feet above the ground, and to-day the stump is very difficult to climb, as the scaffold built around it has broken away.

A trip was made from the General Grant National Park to the upper part of the Comstock millsite, known also as Wigger's. The stage road goes near it and the point is known as "Big Stump." The stump, easily seen from the road, is some 25 feet in diameter with a raised square in the center. The location is in a side basin close to a small brook. An examination of the rings showed that the tree had grown with the greatest rapidity, as the rings were of enormous size. It was estimated to be 1,500 years old. No sample was taken of it. A trip was also made from the park to visit the General Grant Tree and if possible estimate its age. There is an extensive burnt area on the upper side of the tree in which the rings may be observed. These rings are large, and various estimations of the age of the tree obtained in two different visits gave an average of 2,500 years. Near the General Grant Tree is the stump of what was known as the Centennial Tree. It was said that a section of this tree was exhibited in 1876. Since then the stump has been badly burned and is in poor condition for cutting a sample. Some estimate of the rings showed their size to be large, and the age of the tree, therefore, was not very great, perhaps 1,800 years. This confirms the estimate of the General Grant Tree near by.

Table 5. — Sequoia list.

S. No.	H. No.	Huntington's first year of tree.	Identified central complete ring.	First complete ring not central.	Es. loss in cen.	Es. yrs	Probable center of tree.	Prob. age	Len. of rec.	Location.
					cm.			yrs.	yrs.
D1		592 A. D.						1323	1323	Camp 6, Uplands.
2				274 B.C.	4.5	15	289 B.C.	2204	2189	Do.
3			310 B.C.					2225	2225	Do.
4			425 A.D.					1490	1490	Do.
5			1202 A.D.					713	713	Do.
6[5]			1125 A.D.[6]					790[6]	790[6]	Camp 6, Basin.
7				594 A.D.	13.5	75	519 A.D.	1396	1321	Do.
8			294 B.C.					2209	2209	Do.
9			311 A.D.					1604	1604	Do.
10			362 A.D.					1553	1553	Do.
11[5]				697 A.D.	7.5	46		1264	1218	Do.
12[5]	92	17 A.D.		135 A.D.	14	65		1845	1780	Indian Basin.
13[5]	91	585 A.D.	588 A.D.[6]					1327[7]	1327[7]
14	96	387 A.D.	388 A.D.					1527	1527	Do.
15	59	121 A.D.	160 B.C.					2075	2075	Do.
16			506 B.C.					2421	2421	Camp 7, Uplands.
17			308 B.C.					2223	1438[7]	Camp 7, Basin.
18			294 B.C.					2209	2209	Camp 6, Uplands.
19				242 B.C.	8	35	277 B.C.	2192	2157	Do.
20			902 B.C.					2817	2817	Converse Hoist.
21				1305 B.C.	2.3	12	1317 B.C.	3232	3220	Do.
22				1087 B.C.	12.0	75	1162 B.C.	3077	3002	Enterprise.
23				1122 B.C.	14	80	1202 B.C.	3117	3037	Do.

1. The Flagstaff 500-year record.—Figure 3 shows the Flagstaff tree record from 1385 to 1906 A. D., a period of 522 years. The table of measures from which the curve was plotted will be found on page 112. To give the record from 1503 to 1906, 5 trees are used, and comparisons showed that these 5 gave as accurate a record as a larger number whose inclusion would have shortened the record or made awkward breaks in it. The earlier part of the record is from 2 trees
2. Schwappach. Zeitschrift Forst- und Jagdwesen, September, 1904. A recent bulletin of the Mellon Institute, by J. F. Clavenger, entitled, "Effect of the soot in smoke on vegetation," suggests at least a possibility. Clavenger shows photographs of tree sections in the neighborhood of iron mills, in which the growth is normal until the smoke from the mills pours over the forest, and then the rings rapidly decrease in size. It would solve the puzzle if it could be shown that smoke from the iron and brass works in the neighboring city came over the forest of Eberswalde more abundantly at about this time. Dr. Schwappach writes that the manufactories and repair shops are 3 km. distant and in his opinion the decrease in growth can not be due to smoke from them.
3. A letter was published in the local newspaper, asking if anyone had any information regarding forest fires in 1770 or 1821 or of lumbering in that locality in 1807 or 1808, but no reply has been received.
4. Since the trip of 1919 the identification of No. 6 has been fully established.
5. Omitted from the means on account of some deficiency in identification.
6. Identification very nearly right.
7. Not identified after 1130 A. D.

Leaving the vicinity of the General Grant National Park and going south to Porterville, thence by rail to Springville, a 3 days' trip was arranged to the old Enterprise millsite. Camp was made at the cabins, about 10 minutes' walk below the millsite. On going up from the camp. No. 23, known as the Centennial Stump, was found at once, as it is of enormous size, high in the center, and covered with names of visitors. It is located close beside the road and near the wash, about 100 yards from the clear space once occupied by the Enterprise Mill. The oldest tree which Huntington found at this locality had been

	DOUGLASS ⁠ PLATE 7
A
B

A. Cutting sample from stump, Enterprise: D-22, age 3000 years.

B. Centennial stump, Enterprise, Cut in 1874: D-23, age 3075 years.

numbered 116 in his lists. This stump had no number on it, but from the date of its cutting and its age of nearly 3,100 years, it is without doubt the one he refers to. The tree was cut in the winter of 1874-75 for exhibition at the Centennial. The trunk was hollowed out and prepared for transportation in pieces to Philadelphia, where it was said to have been erected, making a sort of hut. In consequence of the uneven surface left, it was very difficult to cut a sample from this stump. However, one was at last secured, which is 12 feet long as it lies on the table in the laboratory.

No. 22 was Huntington's No. 195 and grew near the center of the millsite. Its cutting was extremely easy and its cross-identification with No. 23 and the other trees farther north proved entirely reliable.

The location from which these two interesting trees were obtained is at the very top of a ridge with a steep descent on the east to the North Fork of the Middle Fork of the Tule River and a similar descent on the west to the Tule Valley. The top of the ridge is several hundred yards wide, with opportunity for considerable snow to collect there in winter. It receives little drainage from any source. Just north of it is Mount Moses, high and rugged, and to the south are high ridges extending toward Bear Valley. All the sections obtained in these various trips were shipped to Tucson, and four weeks of continuous work were spent in cross-identification. All the identifications were satisfactory except the year 1580, which was finally determined by the special trip in 1919. The general method of measuring and marking these sections will be found in the next chapter and the tabulation of averages at the back of the book. Owing to the interest in these trees of remarkable size and age, a list of the 23 collected in these two trips is given in table 5.