Journal of Heredity 8(7): 305-314 (July, 1922)

College of Agriculture, University of Illinois, Urbana, Illinois

FIGURE 7. In 1898 Mr. Endicott was successful in obtaining five chestnuts by crossing the Japanese and American species. From these he raised three trees, of which this is the largest. The trees were remarkably vigorous and fruited precociously, one at seventeen months, one at four years, and the other at five years of age. The Japanese chestnut does not usually begin to bear until it is six years old, the American at about twelve years. See text, p. 306.

WITH the rapid disappearance from the American continent of the native chestnut tree through the ravages of the chestnut bark disease, there has come the demand for a substitute. The discovery in China of a species of chestnut resistant to this disease, and the ease with which hybrids between the different species can he produced has led to the hope that hybrid chestnut trees may take the place of the native American species. Mr. J. F. Rock, the agricultural explorer, is now in China collecting all the species of chestnuts he can find there in order that this breeding material may be placed in the hands of American plantsmen, to be used for crossing with the native species to develop the new hybrid chestnut. This article, which calls attention to what a single plant breeder was able to do with the material at his disposal, has a direct hearing on this important problem.

In many parts of the Appalachian Mountains the chestnut is, or was, the most abundant and important forest tree. In some places it made up fifty per cent of the hardwood timber; and it has been estimated to have constituted at least ten per cent of all the hardwood forests in the eastern United States, before the ravages of the bark disease began. Unless it is possible to develop a resistant hybrid that can be easily propagated, one of our most beautiful and useful trees is doomed to disappear from American forests. EDITOR.

1 Paper No. 19 from the Genetics Laboratory, College of Agriculture, University of Illinois. The writers are indebted to Professor J. C. Blair, who furnished some of the facts relating to the early history of this orchard and also make possible an opportunity to visit it.

OCCASIONALLY on some remote farm or ranch a genuine Mendelian experiment of great interest and significance is performed and lies entirely buried until a fortunate accident brings it to light. Our former Colleague, Professor E. W. Bailey, kindly drew our attention to such a case—an orchard of unique chestnut species hybrids. He later placed at our disposal some of the materials which he had collected. We are presenting some of the more obvious and striking facts for the sake of their scientific, historic and practical value. The case is especially valuable, since it affords an excellent example of genetic laws in tree hybridization. In 1899 Mr. George W. Endicott, of Villa Ridge. Illinois, crossed the Japanese chestnut. Castanea japonica (probably variety Coe, but known locally as Japan Giant) with pollen from the American sweet (C. americana). The mere cross in itself was hardly remarkable, inasmuch as several chestnut species crosses had been made before. But Mr. Endicott raised an orchard of about 175 second generation hybrids which are particularly interesting to the geneticist, since all of these trees are now twelve to fifteen years old and are bearing.

Introduction of European and Japanese Species

Edible chestnuts are sometimes divided into three large groups (or species), the American, European, and the Japanese. All three are now grown in this country, but only the first is indigenous. The history of the earliest introduction of the European species is rather obscure, but the general dissemination of this chestnut certainly dates back to about 1800,2 when Eleuthere Irénée du Pont de Nemours planted a number of French chestnuts in his garden near Wilmington, Delaware, where he had settled after emigrating with his family to America in 1799. The descendants of these trees or their scions have been extensively propagated in this general region of the United States. However, there must have been isolated cases of importation before this time, because there is an obscure record3 that Thomas Jefferson grafted the French marron to the American sweet in 1773, at his home in Monticello, Virginia. The importation of the Japanese species was much more recent and was the direct result of attempts to introduce this type on a commercial basis. The S. B. Parsons Co. of Flushing, N. Y., in 1876; William Parry, of Parry, N. J., in 1882; and Luther Burbank, of Santa Rosa, Cal., in 1886, were among the earliest nurserymen and horticulturists to attempt this importation and introduction. The last of these importers planted over 10,000 Japanese seedlings and selected three as desirable—the Coe, the Hale, and the McFarland. Mr. Burbank had several hundred hybrid chestnuts just beginning to bear about 1899—from crosses involving the chinquapin, the Japanese, European, and Chinese chestnuts, and other types, according to Powell's2 report. Van Fleet4 produced hybrids involving European, Asiatic and American types between 1894 and 1911. We have, however, little or no information on the segregation in the second hybrid generation in either of these cases. Mr. Endicott made his first successful cross about this time, i.e., in 1899.

2POWELL, G. H. 11th Annual Report, Delaware Agricultural Experiment Station. 1900.
3BAILEY, L. H. Standard Cyclopedia of Horticulture, MacMillan Co., N. Y. Vol. II. p.742. 1914.
4VAN FLEET, W. Journal of Heredity. Vol. V, No. 1, pp. 19-24. 1914.

FIGURE 8. The results of Mr. Endicott's first crosses were so encouraging that he raised a second generation from the Boone, which he considered the most promising of his three original hybrid trees. As a means of increasing his stock of superior trees this effort was a failure, for the greatest variation was found in the second generation trees. Some were vigorous growers while others were of dwarf habit. Soil conditions may have been a factor in producing the contrast in size shown above, but it is certain that the hereditary constitution of the two trees had something to do with it, as the smaller one is perfectly healthy and bears heavily.
FIGURE 9. All nuts from the same tree are remarkably similar, but those from different trees vary greatly in size and shape and in the amount of tomentum. Some trees have several nuts to the burr, others only one. In some cases the nuts are free in the burr, like the Japanese, while other trees have nuts as firmly attached to the burr as the American sweet.

FIGURE 10. Some of the segregates gave as many as seven or eight nuts to the burr, whereas the Japanese parent has one and the American parent three. The nuts from such burrs are too small and misshapen to be of any great value.
FIGURE 11. The topmost row consists of two Japanese chestnuts on the left and three American sweet on the right. The, next two rows are nuts from the three first generation hybrid trees, the group of four nuts on the left being from the Blair tree, the center four from the Boone, and the four on the right from the Riehl. Note the similarity in size, shape and the amount of tomentum. The last three rows are made up of selections from the 175 trees raised from the Boone, tree, and show the extreme variation encountered in the, second hybrid generation. See text, p. 313.
FIGURE 12. These nuts show further variations in the second generation. The three on the left are acorn-shaped, being from trees that produced but a single nut in a burr. The two on the right are nearly the same size, but they are flat on one side and the amount of tomentum is very much less on one than on the other. See text, p. 314.

Mr. Endicott's Crosses

Mr. Endicott was much interested in hybridization and recognized its possibilities for the chestnut, since he had perfected the Endicott plum and other fruits. Practically throughout his whole life he was interested in plant improvement. Born in 1839, he was 60 years old when he made his first successful chestnut species cross. He was 70 years old when he planted a large orchard of over 150 F2 trees. He died in 1914 without seeing the final results of the second generation of his interesting cross. In choosing the Japanese and the American sweet for hybridization, Mr. Endicott probably had in mind the combination of certain very desirable characters from each parent. He had a large number of American sweet chestnuts growing on his land, but he wished to improve them, for he recognized their shortcomings as well as their desirable qualities. The Japanese type has the advantages of a large nut lying free in a relatively smaller burr, is rather resistant to weevils, blooms and matures early, and has a nut with more attractive color and less tomentum; while the American type has three small nuts very tight in the burr, is susceptible to weevils, blooms and matures later, and has nuts of poor color and heavy tomentum. The American sweet, however, has nuts of fine quality with a thin skin, grows vigorously and produces a large tree, whereas the Japanese has poorer quality (at least in many cases), a thick skin, and produces a low bushy tree of delicate growth. A combination of the desirable features from these two sources naturally recommended itself to Mr. Endicott and therefore he attempted to make the cross.

He seemed to have some difficulty in the routine technique, for the Japanese and the American sweet differed in time of blooming. Eventually he produced five hybrid seeds from which he raised three trees, naming them the Blair, the Boone, and the Riehl. As we might expect, the three trees of the first generation were not all exactly alike, for the parents (at least the Japanese) were more or less heterozygous. As a matter of fact, it is surprising that the three trees were as nearly similar as we found them. Like the American sweet, the Blair and Boone produced three nuts to the burr, while the Riehl produced a single perfect nut with an aborted nut on each side like the Japanese parent. All three first generation hybrids produced nuts free in the burr like the Japanese parent. All of these trees showed tremendous vigor (Fig. 7.) The largest, the Riehl, had a spread of forty-five feet when twenty years old. The Blair and Riehl began to bear at four and five years, respectively, while the Boone bore its first nuts at seventeen months. When we compare this with the Japanese and the American parents, which begin to bear at about six and twelve years, respectively, we gain some idea of the precocity which accompanied the hybrid condition. Van Fleet4 reports similar vigor in crossing Asiatic and European chestnuts with the chinquapin and the American sweet—the Japanese hybrids being again the most precocious. These hybrids were likewise heavy bearers, the Boone producing as much as six bushels in a single season. The vigorous growth of the hybrids is shown by their having a spread of over forty-five feet, while the American sweet and the Japanese parents measured about thirty and sixteen feet, respectively.

FIGURE 13. The American sweet on the left is almost indistinguishable from the adjoining burr, which is a second generation segregate. The burr on the right is also from a second generation tree, but it almost exactly resembles its Japanese ancestor. Only one nut has developed and the two aborted side nuts, or "side blasts," offer no hindrance to the formation of a rounded, acorn-like nut. It is interesting to note that one desirable quality has been passed on by the Japanese parent to nearly all of its offspring. This is resistance to attacks by weevils, whose love for the American sweet has given rise to that ancient "chestnut," "what is worse than finding a worm in a chestnut?" With one exception all of the trees of the Japanese-American cross are free from weevils. Year after year, this tree, although surrounded by an orchard of trees untouched by weevils, is badly infested and harbors a large crop of them. See text, p. 314.

The Second Generation

The first generation hybrids, particularly the Boone with its vigorous growth, early maturity, and with an abundance of nuts of excellent quality and large size aroused Mr. Endicott's interest to such an extent that he planted over 175 seedlings from this particular hybrid tree. A first planting of twenty-five second generation seedlings was made in 1906, and a second planting of over 150 was made in 1909. All of these are undoubtedly genuine second generation hybrids, because the three first generation trees were quite isolated in a field almost a mile from the nearest American sweet. While all the second generation hybrids were grown from the Boone nuts, it is of course possible that the male parent was occasionally either Blair or Riehl, since the three trees were planted in a row and spaced about one hundred yards apart. This does not vitiate our Mendelian experiment, for in any event the second generation trees came from the first generation mated inter se, and probably all or nearly all of them actually came from the self-fertilized Boone hybrid.

In perpetuating the Boone chestnut, by growing a second generation, Mr. Endicott had hoped to see a fairly constant repetition of the excellent qualities of his first three hybrids, namely three large nuts of good color and quality with little tomentum, lying free in the burr, and a vigorous heavy yielding tree. Of course, he was doomed to disappointment in this respect, because the most pronounced segregation was inevitably predetermined. The second generation trees were very uneven in growth and size. The smallest was hardly more than eight feet high, while the largest was about twenty feet high, when they both were fourteen years old (Fig. 8). None of the trees showed the extreme vigor or precocity or heavy-bearing qualities characteristic of the first generation. While the original Boone tree bore its first nuts at seventeen months, the second generation Boone trees were from five to nine years old before bearing; this range almost covers the difference between the original Japanese and American sweet parents. The time of ripening also showed wide variation, since it overlapped both parents. In 1920, for example, the second generation trees ripened their nuts from the first week in September through the middle of October.

In size of nuts the trees showed unmistakable segregation. The nuts of the original hybrid were intermediate between both parents, but these trees gave a greatly increased range of sizes (Fig. 11). Some were as small as the American sweet, and all grades were found up to forms larger than the first generation Boone and almost as large as the Japanese. The amount of tomentum was intermediate in the first generation, and showed a great range in the second. Almost any size nut could be found with any degree of tomentum. The first two and last two F2 nuts in Fig. 11 illustrate this point. While the nuts on different trees showed a great range of variation in size, form and tomentum, the nuts on any individual tree were remarkably uniform. Figure 9 shows two groups of these second generation nuts. Each group, coming from a single tree, is uniform within itself, but one group gives consistently larger nuts, with slightly more tomentum.

Most of the trees gave three nuts to a burr, but some gave only one nut with two aborted nuts like the Japanese parent (Fig. 13). and others gave as many as eight nuts in a burr (Fig. 10). When a tree produced singles these were often ovoid or acorn-shaped (Fig. 12). While all three of the original hybrid trees produced nuts free in the burr, segregates appeared in the second generation with nuts just as firm as those of the American sweet. Probably fifty per Scent. of the trees produced free nuts, while six per cent. were very tight, the remainder being intermediate. Segregates closely resembling the American sweet were found, both in size, shape, number of nuts and character of burr (Fig. 13).

Resistance to Weevils

For many years all the trees of both generations with one exception have been resistant to weevils like the Japanese parent. The single exceptional tree has always given nuts badly infected with weevils, while all the rest of the trees in the orchard have been immune.

The character of the burr also showed much variability in the second generation in respect to thickness, length of spines and the like. The spines of some burrs were relatively soft and easily handled, while others were extremely rigid. All gradations between these extremes were found.

When it was found that the second hybrid generation was so exceedingly variable and not as valuable as the original hybrid, an attempt was made to propagate the parent vegetatively by grafts and by top working the worthless second generation seedlings with original Boone wood. Mr. Endicott was accomplished in the art, but never had much success in the use of F1 wood on F2 trees. In many trials (over 400) made by Mr. Endicott and Mr. E. A. Riehl (an expert nurseryman) about three per cent. of the attempts to bud or graft were successful. Possibly some obscure anatomical or physiological peculiarities of the wood of the first generation make a union with other woods difficult, even with its own seedlings.

All of these striking variations in the second generation trees bear the earmarks of multiple factor segregation and recombination. There are many other characters involved besides these more patent examples which we have chosen to record. No doubt intricate structural and physiological characters are also included. The orchard might prove to be a storehouse of promising material for anyone inclined to pursue investigations in this direction.