UNIVERSITY OF CALIFORNIA PUBLICATIONS IN AGRICULTURAL SCIENCES Vol. 2, No. 1, pp. 1-46, pis. 1-12 December 4, 1913 STUDIES IN JUGLANS I STUDY OF A NEW FORM OF JUGLANS CALIFORNICA WATSON BY ERNEST B. BABCOCK CONTENTS PAGE I. History and description. First appearance and description of the original trees 2-5 General appearance - 3 Leaves — - 4 Nuts 4-5 Recurrence of the new form 5-6 Distribution under misleading names 7 Name and botanical description 7-8 Comparison with the species 8 Summary 9 II. Origin of the new form — hypotheses, observations, and experiments. Two discarded hypotheses 9-11 First working hypothesis — hybridization 11-16 Results of hybridization experiments 15 Second working hypothesis — teratology 16-20 Summary of evidence 19-20 Third working hypothesis — mutation 21-23 Summary of evidence 23 2 University of California Publications in Agricultural Sciences [Vol. 2 STUDY OF A NEW FORM OF JUGLANS CALIFORNICA WATSON I. History and Description In the autumn of 1900 D. C. Disher, of Garden Grove, Cali- fornia, according to his own account, gathered about two thous- and nuts from a certain California black walnut tree that grew near Yorba, in Santa Ana Canon, but which has since been destroyed. These nuts were planted in the spring of 1901 in order to raise seedlings upon which to graft the English walnut for orchard planting. Among the seedlings about twenty appeared from the first to be distinct from the rest, which resembled the parent tree. Of these twenty only two trees remain where they were first transplanted from seed-bed to nursery row, the others having been given away or destroyed. Of these two only one produces both male and female flowers and bears nuts ; the other always produces staminate catkins in abundance but no pistillate flowers. For this reason the first mentioned individual has already been described 1 as "the original fertile tree," but it would have been more exact to have referred to it as Disher 's fertile tree, inasmuch as some of the other original specimens above mentioned have been located and are known to bear nuts also. The writer has seen seven of these distributed trees, and material from an eighth ; in leaf and bark characters as well as in general appearance they resemble Disher 's trees. Three of these are growing at the experimental garden of N. B. Pierce in Santa Ana, three are located in the town of Garden Grove, one is on the Leffingwell Ranch in East Whittier, and one is at the United States Plant Introduction Gardens in Chico, Cali- fornia. One of Pierce's trees is shown in plate 13, figure 16. The discoverer of these trees wished to preserve them because they are so strikingly different in their leaf characters and in general habit from ordinary California black walnut. They pos- sess no special economic value, being less robust than other wal- nuts and more restricted in their range of adaptability to adverse i Babcock, E. B. t in Jepson, The Silva of California, Mem. Univ. Calif. II (1910), pp. 50-54. ]913] Babcock: New Form of Juglans 3 soil conditions. But their structural characteristics alone are sufficient to excite the interest of the student of plants, especially of one interested in problems of heredity and evolution. Perhaps the most interesting thing about these trees is their resemblance to oak trees. In mass effect they resemble small- leaved oaks more than walnuts. This is mostly due to the small size of the leaves and to their color, which is a darker or duller shade of green than that of California black walnut leaves. These features, associated with the fact, noted by Disher, that the parent tree stood close beside a coast live oak tree (Quercus agrifolia Nee), are sufficient to account for the view, held by a number of persons, that this new form originated through hybridization between walnut and oak. As the seeds were planted in 1901, these trees are now twelve years old. The two retained by Mr. Disher were left in the nursery, which was set out to commercial varieties of walnut {Juglans regia) later on by the owner of the farm. He has allowed the trees to stand unmolested, except for trim- ming up low-hanging branches. Now they have attained a height of twenty-five feet and have a spread of branches about twenty feet in diameter. Seen among the broad-leaved English walnuts, these two trees present a distinct appearance with their many slender branchlets and their sparse foliage. In early spring and late autumn or early winter the contrast is even greater, because these trees and other specimens of the new form resemble south- ern California black walnuts in the brevity of their dormant period. They leaf out very early in spring and some leaves per- sist until February. The English walnuts, however, come into leaf from April to June, according to the variety, and by Nov- ember are once more leafless. Other distinctive vegetative characters are well shown in plate 1, figure 1. There is a marked tendency to dichotomous branch- ing. This is conspicuous in the left-hand tree in the picture and is noticeable in the other individual. There is also a tendency to form bunches of leaves at the ends of the branchlets. This is apparent in both trees. The appearance of the bark on the trunks of Disher 's original trees is also distinct from that of Juglans calif ornica. "While that of the latter in trees of the 4 University of California Publications in Agricultural Sciences [Vol. 2 same age as Disher's is usually rough, the bark on these trees is rather smooth. The left-hand tree in plate 1, figure 1, always produces both male and female flowers and frequently bears abnormal bisexual flowers also. But the other tree produces only staminate cat- kins, and the photograph shows in what profusion they are borne. A very few pistillate flowers have been seen on this tree, but it has never been known to bear fertile seeds. That sterility in this individual is associated with greater vigor in wood production is shown in the above figure. The leaves on some specimens of the new form are highly variable. The common type of leaf, however, is one with a terminal leaflet one or two inches long and two smaller lateral leaflets, as shown in plate 2, figure 2, a. Sometimes both lateral leaflets are missing and occasionally only one is present, as shown in plate 2, figures 2, b and c, and in plate 3, figures 4 and 5. Plate 2, figure 2, is approximately natural size. In order to appreciate the remarkable difference in leaf characters between the new form and Juglans californica note the leaves in plate 7, figure 11. The frequent occurrence of ascidia (pitchers) on the leaves of the Leffingwell "original" tree gives further evidence of the tendency toward extreme variation. Not being aware that ascidia have been reported previously in the genus Juglans, I give plate 2, figure 3 a photograph of four leaves bearing ascidia and one normal leaf. The nuts borne on Disher's original fertile tree resemble those of southern California black walnuts in external appearance of the husk (cf. plate 2, figure 2). Their average size is much smaller. The immature fruits are also similar, except that many of them are ridged or grooved as in plate 3, figure 4, a, and figure 5, a, while some depart widely in appearance from the typical young fruit of Juglans californica. Such a specimen is shown in plate 3, figure 5, b. The smooth area surrounding the stigma was proportionately large, and tapering arms extended towards the base. Surrounding the stigma were several small protuberances, apparently remnants of anthers. The occurrence of bisexual flowers on this tree has been mentioned. When the dry husk is removed from the mature nuts of this tree, they are 1913] Babeock: New Form of Juglans 5 seldom found to be divided into nearly equal parts by a deep suture but there are sometimes outer indications of tripartite inner structure. Such marks are shown in plate 4, figure 6, I. The occurrence of tricotyly is frequent but many tripartite nuts con- tain no embryos. Such a nut containing an embryo is shown in plate 4, figure 6, k. In dicotyledonous nuts from this tree the cotyledons are reduced, sometimes so much so as to be barely distinguishable (cf. plate 4, figures g, h, i, j). Recurrence op the New Form In the autumn of 1907, when the writer first visited Garden Grove to examine the original trees, he was shown about a dozen two-year-old seedlings, all of which closely resembled the original trees. These were found scattered through the nursery of about nine thousand budded English walnut trees. The seedling roots were grown from nuts collected in the autumn of 1904, partly from wild trees in Brea Canon and partly from a row of Juglans calif ornica trees growing in Garden Grove. The seeds had been mixed so that there was no way of locating the tree or trees that gave rise to the new form. Mr. Disher's interest in the unusual appearance of these seedlings again prompted him to leave them unharmed, so that the writer was able to secure a fairly good photograph of one that stood at the end of a row (cf. plate 5, figure 7). The contrast between this little tree and its luxuriant neighbors was certainly striking. The slender branches and ovate leaflets are distinctive. Some leaves are so placed that the large terminal leaflet and two small lateral leaflets are clearly shown. One of the seedlings differed from all the rest in having two pairs of lateral leaflets. Nearly all the leaves on the tree were of this type (cf. plate 5, figure 8). Since that time the writer has seen similar leaves on other seedlings, most of whose leaves were three-parted, as well as leaves intermediate between the two. In fact, the great amount of variation in the leaves of the several individuals which we class as the "new form" is one of the most interesting things connected with it (cf. plates 2, 3, 5, and 11). In 1909 William Tyler, son-in-law of Mr. Disher, reported to the writer that he had found a few specimens of the aberrant 6 University of California Publications in Agricultural Sciences [Vol. 2 form among a large number of seedlings of the Garden Grove black walnuts. This was the third appearance of the form in question. In the autumn of 1910, the writer arranged to have the nuts from the various individual black walnut trees of the row in Garden Grove gathered separately so that they might be planted separately the following spring. The nuts were so gathered, but unfortunately became mixed while in storage. Thus, although about thirty specimens of the new form appeared, they were scat- tered through the nursery and could be traced to no particular tree or trees. This was the fourth appearance of the new form. In the autumn of 1911 other parties secured many of the nuts from the Garden Grove trees and although a few aberrant seedlings appeared in Tyler's seed beds, he was not certain which tree produced them. This makes the fifth appearance of the form . In 1912 the writer had the product of twenty-one of the trees in the Garden Grove row gathered separately. Among the sprouted seedlings of one tree six aberrant seedlings have already appeared, making the sixth appearance. An additional appear- ance of the new form has been reported. In this case a single seedling appeared among those grown from a mixed lot of south- ern California black walnut seeds from trees growing wild in Santa Monica Canon. The seeds were gathered in 1910, so that the tree is now three years old. It resembles Disher's original trees. From the foregoing account it is evident that the new form has originated in at least three different localities. In two of these, Santa Ana and Santa Monica canons, the trees were in the wild condition when the nuts were collected. The Garden Grove trees comprise a boundary line planting between two farms. The new form is not reported to occur in the wild and probably does not so occur. However, it would no doubt thrive in the more favorable areas now occupied by wild walnuts in southern Cali- fornia. Seedlings of Disher's original trees have been observed to suffer more from excess of moisture than from drouth. Walnut breeders and nurserymen have propagated the form by means of grafting. Frank A. Leib, of San Jose, has a 1913] Babcock: New Form of Juglans 7 young tree grown from a cion obtained from N. B. Pierce, of Santa Ana, and grafted on a hybrid walnut root of the "royal" type. The tree has made rather remarkable growth. The Gar- dena Agricultural High School has obtained seeds of Disher's original tree at Garden Grove in order to grow seedlings for instructional purposes. But there is general confusion as to the true nature of the form and it has been distributed under mis- leading names. In view of these facts it was deemed advisable to record a name indicating its natural relationship preliminary to the publication of this paper. 2 Now, if this form had been first described from specimens collected in the wild, there is no doubt that it would have been named a distinct species. Without endeavoring to solve the prob- lem of its origin the botanist would have felt justified in thus naming it by its absolutely distinct foliage alone. But we do not know that it could long exist under wild conditions and, as will be shown later, the seedlings of the original trees do not come uniformly true to type. Instead of so doing, usually some of them resemble J. calif ornica. Hence, the writer feels justified in calling it a variety of the species from which it has sprung, no matter by what process. The following is a description of Disher's original fertile tree already published 3 with two or three minor changes based upon data in the possession of the writer. NEW VARIETY JUGLANS CALIFORNICA var. QUERCINA Babcock Tree 20 ft. or more high. Bark aromatic and strongly walnut- scented. Branchlets hollow, chambered with pithy plates. Twigs, bud- scales, and young leaves granular-pubescent. Buds few-scaled, axillary or superposed. Leaves 1 to 3 inches long, alternate, exstipulate, mostly compound with three leaflets; terminal leaflet two or three times as long as lateral leaflets and ranging from % to 2 inches in length, in form varying from broadly ovate through oval to elliptical or oblong, truncate or emarginate at the apex, margin serrate or almost entire; lateral leaflets placed opposite or scattered, with petiolules or sessile, sometimes one or both lacking; petiole equal to, shorter or longer than, the terminal leaflet. Plant monoecious, occasionally with hermaphrodite -Babcock, 'A New Variety of Juglans calif ornica Wats." Science, n. s., XXXVIII, 968, p. 89. s Babcock, Mem. Univ. Calif. II (1910), p. 54. 8 University of California Publications in Agricultural Sciences [Vol. 2 flowers. Staminate flowers in lateral catkins from wood of the preceding year; calyx adnate to the inconspicuous dark-red bract, irregular, con- sisting of three larger and one or two smaller lobes and an inner whorl of 4 to 6 smaller distinct sepals, one or two of which sometimes show stamen characters; stamens sepaloid, 10 to 13, with 1 to 4 of the central ones abortive; filaments free, very short; anthers variable, the pollen sacs unequal, especially in the outermost stamens, connective not bifid at the top. Pistillate catkins 3 to 6-flowered, terminating branchlets of the same season's growth; calyx irregularly 4 to 7-lobed, adherent to the inferior 1-celled ovary; the latter often with 1 or 2 longitudinal grooves or ridges, rarely with 2 or 3 fleshy bracts near the base, occasionally with abortive anthers; styles 3 or 4, short, united toward the base oi free, stigmatic along the inside, the fringed surfaces forming a rosette. Fruit similar in appearance to that of Juglans californica Wats., but smaller and more variable as to form and internal structure of the nut; the seed also much smaller; cotyledons much reduced, not convoluted (cf. plate 1, figure 1, plate 2, figures 2 and 3, plate 3, figures 4 and 5, plate 4, figure 6, g, h, i, j, k, I, plate 5. figures 7 and S, plate 6, figure 9, plate 11, figures 17 and 18). Differences between the New Form and the Species Type J. californica Leaves, 6 to 13 inches long, coin- pound with 11-19 leaflets. Terminal leaflet shorter than, equal to, or longer than the lateral leaflets and ranging from 1% to 4 inches long, oblong lance- olate, serrate. Staminate flowers. Calyx irregu- larly 3 to 6-lobed; stamens 20 to 26, connective bifid at the apex. Pistillate flowers. styles 2. Calyx 4-lobsd; Fruit globose, % to 1 inch in diameter; cotyledons promin- ent, much convoluted. J. californica var. quercina Leaves 1 to 3 inches long, mostly compound with 3 leaflets, rarely with 5, sometimes simple. Terminal leaflet two or three times as long as lateral leaflets and ranging from % to 2 inches in length, in form varying from broadly ovate through oval to elliptical or oblong, truncate or emaroinate at the apex, margin serrate or almost entire. Staminate flowers. Calyx adnate to the inconspicuous dark red bract, irregular with 3 larger and 2 smaller lobes and inner whorls of 4-6 smaller distinct sepals; stamens sepaloid, 10 to 13, anthers variable, connec- tive not bifid at apex. Pistillate flowers. Calyx irregular- ly 4 to 7-lobed; styles 3 or 4, short, the fringed surfaces forming a rosette. Fruit smaller and more variable as to form and internal structure; seed much smaller, cotyledons much reduced, not convoluted. 1913] Babcock: New Form of Juglans SUMMARY 1. A new form of walnut has appeared on seven separate occa- sions among seedlings of at least three different trees of Juglans calif ornica Wats. 2. This form is sufficiently distinct from all other walnuts to justify its recognition as a new species. But in all but one of the germination tests of seeds from the original trees, some seeds have produced plants resembling the species type in leaf characters. Moreover, the form is exceedingly variable. These facts are good reasons for describing the new form as a variety of J. calif ornica. II. ORIGIN OF THE NEW FORM— HYPOTHESES, OBSERVATIONS AND EXPERIMENTS In studying the nature and origin of this new form of walnut, three working hypotheses have been retained, and two other hypotheses have been eliminated so far as my experimental work is concerned. The latter hypotheses will be discussed briefly before passing to the fuller consideration of the working hy- potheses and the investigations connected therewith. One of the first possibilities suggested was that the new form may be a "reversion to an ancestral type." The species is long- lived and stump-sprouts freely. A few generations might extend back to a time when a now extinct form existed in the same area. It is a well-known fact that our present American species inter- breed freely when growing near each other, and also breed with Juglans regia under similar conditions. Hence it is conceivable that such an extinct form might have interbred with J. cali- f ornica. According to Mendelian principles, some of the progeny would be heterozygous for certain characters, which might be the distinguishing characters of the new form. Such heterozygous progeny, under favorable conditions, would continue to produce both parent forms and more heterozygous individuals. It is conceivable, then, that the few trees thus far known to exist, which are giving rise to the new form, are such heterozygous individuals, and that our new form is really a supposed extinct 10 University of California Publications in Agricultural Sciences [Vol. 2 form, whose existence is revealed through segregation of unit characters in the gametes of the heterozygous individuals. But this conception was unsatisfactory as a present working hypo- thesis, since it necessarily assumes a parent form, the existence of which can only be proved by paleontological records. So far as the writer has been able to ascertain, none of the extinct species of Juglans or Carya thus far described resemble the new form in number and shape of leaflets and proportionate size of lateral and terminal leaflets. 4 Herbarium material of other related genera has also been examined to see whether a suggestion of the new form's leaf characters could be found, but without suc- cess. However, until some other hypothesis is proved to fit the case, this one should be reserved as having some value. Another possibility is that the new form is a hybrid between Juglans calif ornica and some other species, such as J. regia or J. nigra. This suggestion has not been used as a working hypoth- esis because the characters of hybrids between these two species and J. calif ornica are already known. It is generally understood that plants of the F 1 generation of both these crosses reveal the partial dominance of J. nigra or J. regia as the case may be. 5 They always have larger leaves than those of J. calif ornica, whereas our new form is characterized by its small leaves. It is hardly conceivable, then, that the new form is the direct result of such a cross, nor even that the wild trees from which our new form springs could belong to the F 1 or a later generation from such a cross and thus produce the new form among the extreme variations that sometimes occur among the younger generations of hybrids between species. For our new form is the only extreme variation which has been reported among the many thousands of J. calif ornica seedlings that are grown annually in California. Moreover, <7. regia has been cultivated in southern California for less than fifty years, while only a few cultivated trees of J. nigra occur, and these also came with the introduction of commercial walnut growing. Other known species of Juglans * Dr. W. A. Berry, in a letter to the writer, expressed the following opinion: "I suspect if a paleobotanist had come across such a form, and I know of none such, he would have thought of the Anacardiaceae rather than Juglans. ' ' s Smith, R. E., "Walnut Culture in California," Univ. Calif. Agr. Exp. Sta. Bull. 231, pp. 157-170. 1913] Babcock: New Form of Juglans 11 are also sufficiently distinct from J. californica so that hybrid offspring of the F 1 generation would hardly resemble our new form. Exceptions to this would be J. rupestris and J. major, the Texan and Arizonan species, which resemble J. californica rather closely, and on this account the possibility of our new form being an F 1 hybrid between either of these species and J. califor- nica is even less likely. Hence, while it is very desirable that a systematic study of hybridization among all species of Juglans should be made, because of its botanical, genetic and horticultural interest, the writer does not consider it likely that the new form is a hybrid between J. californica and any other known member of the genus. By analogy we may eliminate all other members of the family Juglandaceae from similar participation in the origin of the new form. Turning next to the three hypotheses which serve as a basis for the investigation now in progress, it would seem that sufficient work along any one of these lines might lead to a solution of the problem. They will be discussed in the order in which they were originally taken up by the writer. They are as follows : (1) The new form may be a natural hybrid between Juglans californica Wats., and Quercus agrifolia Nee, or some other oak. (2) The new form may originate in certain teratological flowers that have been discovered on certain Juglans californica trees. (3) The new form may be the result of mutations in certain male or female flowers (or both) of certain Juglans californica trees. First Hypothesis Let us consider first the possibility of origin through hybrid- ization between walnut and oak. The original trees were first shown to the writer as "crosses between a walnut and an oak." They were briefly described by him under the title ' ' The Walnut- Oak Hybrids," in Jepson's The Silva of California (pp. 50-54). In 1907, N. B. Pierce verbally expressed to the writer his opinion that the new form is a hybrid. Since the conception of hybrid origin was entertained by various persons, it seemed wise to investigate the possibility, not only of the occurrence of natural 12 University of California Publications in Agricultural Sciences [Vol. 2 hybrids, but also of producing such hybrids artificially. From the descriptions of the two families it is evident that they are rather closely similar. Following is a table of comparison" of the reproductive organs in Juglans and Quercus : Juglans Plants monoecius. Staminate flowers on lateral pen- dulous catkins on last season 's wood; calyx irregularly 3 to 6-lobed; stamens numerous. Pistillate flowers solitary or few in a short terminal spike; calyx 4-lobed, adherent to the 1- celled inferior ovarv; styles 2. Fruit a 1-eelled, incompletely par- titioned nut, 1-seeded, the seed so lobed as to fit the irregulari- ties of the cavity, exterior of nut covered with green and fleshy or at length dry and brown husk. Quercus Plants monoecius. Staminate flowers on pendulous (ex- cept in one species) catkins from buds of the previous sea- son; calyx parted into several lobes; stamens 4 to 12. Pistillate flower 1 in an involucre; involucres 1 or 2 in the upper axils of the season's shoot: calyx adherent to the 3-celled, 6-ovuled ovary; ovary with 3 to 5 styles or stigmas. Fruit a 1-celled, 1-seeded nut, only 1 ovule maturing; seed with thick, fleshy cotyledons; the nut set in a scaly cup. From the above table it is evident that in gross structure of flowers the two genera are closely similar. Of course, there is always the possibility or likelihood that some feature in the minute structure of the pistillate flowers or in the physiology of fertilization may absolutely prevent hybridization between any and all species of the two genera. The writer has not yet been able to engage in the cytological study necessary to confirm or deny this possibility. The first effort made was to search for chance natural hybrids in the wild. In 1907 about four thousand nuts of Juglans cali- fornica were gathered in Brea Caiion. They were taken from trees standing close to coast live oaks. As they were planted late in the spring they were slow in germinating and cutworms destroyed many of them, but among two hundred that grew during the summer of 1908, no aberrant forms appeared. No further efforts in this direction were made, as the possibilities of success were considered too remote to warrant the expense of collecting and growing. 8 Based on descriptions of the genera in Jepson 's The Silva of Cali fornia. 1913] Babcock: New Form of Juglans 13 In undertaking cross-pollination experiments it was reasoned that the failure of a large number of careful efforts to secure such a hybrid would discredit this hypothesis, while the produc- tion of one such hybrid artificially would tend to strengthen it. During the first year's work the only oak experimented with was the coast live oak, Quercus agrifolia, but in succeeding years one or more other species were included in the trials. The female flowers used were all on J. calif omica trees in 1908, but since that year an effort has been made (yet without success) to secure reciprocal crosses. Experiments in 1908, 1909, and 1910 Two indigenous walnut trees were selected. They were located, one at the rear and the other at the front of a large city lot in the suburbs of Los Angeles. Through the courtesy of the residents, they were protected from interference during the critical stage of the work. They will be designated as Tree I and Tree II. Manila paper bags were placed over the pistillate catkins almost as soon as they appeared and before pollen was being shed by the staminate catkins. Oak pollen was collected in homeo- SUMMARY FOR 1908 Tree I bource of pollen Q. agrifolia .... Number of pistillate catkins pollinated 23 Number of catkins on which nuts formed 14 Number of nuts produced 27 Nut gem in s that linated 1909 26 Trees growing in 1913 24 Freak S. 19 8 13 13 12 Checks 8 Tree II Q. agrifolia .... 17 16 37 33 32 Cheeks 5 1 2 2 2 pathic vials and applied with camel's-hair brushes. In 1908 no pollen was being shed on Tree I, even at the time of pollination. On Tree II pollen was being shed but care was used not to expose 14 University of California Publications in Agricultural Sciences [Vol. 2 the pistillate flowers when pollinating them. The only other source of pollen in 1908 was Disher's original sterile tree, which is referred to below as Freak S. In the following table "checks" are bags that were left in place at time of pollination and allowed to remain for several weeks, in order to learn whether nuts would develop. The trees referred to as growing in 1913 are located on the campus of the University of California. In 1909. the walnuts were very late in blooming and conditions were so adverse that no nuts whatever were produced as a result of pollinating thirty-four different catkins (about sixty flowers). In 1910 conditions were favorable and very promising. In May, 1910, there were 151 nuts developing as a result of pollin- ating 79 pistillate catkins with Quercus agri folia pollen, and 29 nuts as a result of pollinating 16 pistillate catkins with Q. engel- manni pollen. But through a miscalculation on the part of an assistant regarding the proper time to secure the nuts before they dropped from the trees, the entire lot was lost. However, this assistant, S. E. Goodall, made similar experiments at his home near Chatsworth, using pollen from Q. agrifolia and Q. lobata. The oak-pollinated nuts were protected and saved, and as a result there are growing on the campus of the University of Cali- fornia eight young trees from Q. agrifolia pollinations and four from Q. lobata pollinations. Experiments in 1911 The trees used are located on F. Goodall 's ranch near Owens- mouth, Los Angeles County. Three different trees were used. They will be designated by the letters A, B, and C. Coffee bags, having the outer layer of oiled paper, were used to cover pistillate flowers. An effort was also made to secure reciprocal hybrids. One tree of Quercus lobata was found which was shedding but little pollen and which was somewhat removed from others of its kind. Bags were placed over seven shoots of the season's growth, thus covering forty to fifty pistillate flowers, after first having pollinated them with J. calif omica pollen. Three weeks later the bags were removed and some flowers seemed likely to develop further, although some were moldy. On July 21 all had dropped. abcock: New Form o f Juglans IE Summary for 1911 Tree A Number of pistillate catkins pollinated 20 Number of nuts produced 9 Nuts germinated June 1, 1912 20 9 12 4 2 1 1 4 Tree B 24 18 15 13 10 7 1913] Source of pollen Q. agrifolia Q. lobata Q. dumosa Checks Q. agrifolia Cheeks Trees B and C" Q. lobata 44 29 22 Results of Hybridization Expt riments From the 1908 pollinations with Quercus agrifolia upon Jug- lans calif omica there are now 48 four-year-old seedlings. Cions from most of these have been grafted this spring upon large English walnuts in order to hasten fruiting. It is planned to protect these grafts at the blooming period so as to secure self- fertilization and thus make possible the " breaking-up " which may occur in the F 2 generation if they are really hybrids. From the 1910 experiments (made by Goodall) there are eight seedlings from the Q. agrifolia pollinations and four from Q. lobata pollinations. Grafts may be made from these next year. From the 1911 experiments there are 14 seedlings from the Q. agrifolia pollinations, 20 from the Q. lobata, and 2 from the Q. dumosa crosses. The seedlings obtained from all oak pollinations resemble ./. calif omica in leaf characters and habit of growth. This does not necessarily mean that these seedlings are not true hybrids. It has been demonstrated that certain species-hybrids are goneoclinic to the female parent in the F 1 generation. 8 It is in the hope of showing that these seedlings are hybrids by the appearance of oak characters in plants of the F 2 generation that the seedlings are being propagated to secure early fruiting. 9 7 While in storage the Q. lobata nuts from Trees B and C became mixed, but 24 of the 29 came from Tree C. s Keeble, F., The Gardener's Chronicle, vol. 52, no. 1355, p. 457. :) Since writing the above it has been found that most of the seedlings from the 1908 crosses bloomed this spring (1913). Hence it was possible to secure self-pollination on the seedlings themselves at an earlier date than was anticipated. 16 University of California Publications in Agricultural Sciences [Vol. 2 In 1908 the check results were much more satisfactory than in 1911. Various factors may have influenced these results, hut the writer is inclined to think that the comparative isolation of the Los Angeles trees made them more satisfactory subjects for experimentation. From the 1911 results we find nearly as high a percentage of nuts in check bags as in pollinated bags from Tree A and a higher percentage from Tree B. The production of nuts in the check bags raises the question of the possibility of apogamic development of seed in Juglans calif arnica, and throws a shade of doubt upon the supposed hybrid seedlings that have been obtained from oak pollinations. In this connection more experimental work should be done in preventing natural pollina- tion on a large scale. For the present the writer is inclined to think that one of the other two hypotheses is much more likely to reveal the origin of our new form. Dr. Cannon, of the Desert Laboratory, Car- negie Institution of Washington, who is an authority on plant hairs, informs me that he finds no evidence of oak characters in the hairs of Disher's original trees. Second Hypothesis Let us consider next the possibility that the new form may originate in certain teratological flowers on certain Juglans cali- fornica trees. It is necessary to give here a brief account of the discovery of these abnormal flowers and to state the reasons for considering them as a possible source of the new form. In order to emphasize the teratological features about to be described, it may be well to consider first the characteristics of normal fruits and flowers. Normal flowers are shown in plate 6, figure 10. Normal leaves and fruits are shown in plate 7, figure 11. The normal blooming period of most wild black walnut trees in southern California is in April. As the staminate catkins are produced upon the wood of the previous season's growth, their gradual increase in size may be observed at any time during late winter or spring. During March they lengthen noticeably. About April first the pistillate catkins appear, terminating the first new growth of the season. They are one or two inches long and bear one to five flowers, so that, when the fruits mature. 1913] Babcock: New Form of Juglans 17 they often hang- in clusters (cf. plate 7, figure 11). Normal pistil- late flowers are bisymmetrieal. Normal staminate catkins are two to four inches long, pendulous, and bear an indefinite num- ber of flowers (cf. plate 6, figure 10). Many trees throw out lateral branchlets from the first growth of the season. It is usually during this second period of growth, in May or early June, that teratological flowers and leaves appear on certain trees. (It should be noted here that abnormal female flowers have been observed by the writer during the normal blooming period of two or three trees. A note on such specimens with illustrations appears in Jepson's The Silva of California, pp. 55, 56. So far as is known these flowers never produce fruits containing viable seeds.) The late or second-growth abnormal flowers are usually produced upon catkins that resemble normal staminate catkins in number and arrangement of flowers. But the flowers are either pistillate or bisexual, often both occurring on the same catkin. Only a few staminate flowers have been observed among these late blooms, and these were on catkins which were entirely staminate. One tree was observed on May 2!), 1909. on which nearly all the late or second-growth catkins were staminate, but the flowers were dropping instead of devel- oping to maturity. Most late-appearing catkins occur on the second-growth, lateral branchlets, one catkin in the axil of a leaf. but they sometimes develop alongside the normal, terminal, pis- tillate catkin, as shown in plate 8, figure 12. In Brea Canon (Puente Hills) a hundred or more wild trees were examined during the season of 1909, and, while abnormal flowers and fruits were not of general occurrence, they were found to be verj frequent. The most striking characteristic of the pistillate and bisexual flowers is their form. They are asymmetrical, being flattened more or less on the side adjacent to the axis of the catkin. Along with this flattening there is often a depression in the surface of the ovary, usually extending from the styles to the base of the ovary or only part way, and of varying width (cf. plate 8, figure 12). When teratological flowers develop into fruits, the asymmetri- cal form and principal surface markings are retained. Over fifty 18 University of California Publications in Agricultural Sciences [Vol. 2 abnormal fruits, with their husks dry and brown but in most cases uninjured, were picked up by the writer beneath one of the several trees upon which such fruits have been known to mature. But most of these trees produced only a few such fruits. Mature abnormal fruits are easily assorted into two or three lots according to their external markings. Very few differ notably from the two classes illustrated in plate 9, figure 13. However, in each class there are greater extremes of variation than are here shown. The series a, b, c has been designated "Type X" and the other series, "Type Y." In a, b, c, d, e, the fruits appear as when picked up; a, b, c, are seen in longitudinal plane, while d and e are seen from above; in a', b' , c', d' , e' , the dry husk has been entirely removed from the woody shell of the nut. These nuts are from fruits which resemble a, b, c, d, and e so closely as to be practically identical. While a', c' , e' are seen from above, b' and d' are shown in longitudinal plane. The nuts a", b" , c", d", e" are practically identical with a' , b' , c', d' and e' . They were sectioned transversely and placed in relatively the same posi- tions as a', b', c', d' and e' respectively, the basal portion of the nut being placed in the upper of the two rows in each case. To me, the most striking abnormality shown in this picture is tin; unequal reduction of the cotyledons in all except e" . Two culture tests of abnormal nuts selected to these types have been made in 1910 and 1911. There is considerable varia- tion among the seedlings grown from each type, but no general differentiating character could be found among all the seedlings from the two types of nuts. These tests were subsidiary to the general trials of abnormal nuts conducted during the same years in order to ascertain whether the new form originates from such nuts. Among sixty-eight seedlings growing not one has shown the slightest indication of the leaf characters of the new form. In 1910 trials one seedling appeared that had three-parted scale- like leaves, but it died before it was three inches high. The leaves did not resemble those of the new form. On otherwise normal Juglans califomica trees, abnormal leaves have been observed in two situations — near the base of second-growth lateral branchlets and, occasionally, associated with the abnormal catkins already described. Abnormal leaves, from 1913] Babcock: New Form of Juglans 10 second-growth lateral branchlets on trees bearing abnormal flowers, and leaves associated with abnormal catkins are shown in plate 9, figure 14. Some of these leaves resemble somewhat the typical leaves of the new form and, at the same time the leaves shown here were collected, one second-growth lateral branchlet was found that bore a leaf very similar to the typical leaf of the new form (cf. plate 10, figure 15, a). We may now summarize the phenomena observed in connec- tion with the occurrence of teratology in Juglans calif arnica and add some considerations with respect to the possible origin of the new variety. Abnormal flowers, fruits and leaves are of frequent occurrence on indigenous trees. They usually occur later in the growing season than the normal blooming period, on secondary, lateral branchlets, or in the case of abnormal catkins, sometimes as secondary catkins terminating the first growth of the season. The fruits produced by abnormal flowers retain the characters of asymmetry and irregularity of surface. Their average size is only about half that of normal fruits. Only a small percentage of the abnormal fruits collected by the writer contained viable seeds. Among all the seedlings which have been grown from abnormal nuts, not one shows a trace of the leaf characters of the new form. On the other hand, in con- nection with the fourth recurrence of the new form at Garden Grove in the spring of 1911, the writer was able to ascertain definitely that some of the seedlings of the new form grew from nuts of normal size and shape. These facts indicate that the new form does not originate from teratological fruits. However, we must concede the possibility of the original trees having so originated. More of these abnormal nuts should be collected and the seeds tested. The reasons for thinking that the new variety may have orig- inated from teratological fruits may be concisely stated as fol- lows : 1. There is more or less similarity between the abnormal leaves, found on secondary branchlets or associated with abnor- mal catkins, and the leaves of the new form. 2. Abnormal flowers and fruits are frequently found during the normal blooming period on some of the original trees of the 1 I'd University of California Publications in Agricultural Sciences | Vol. :_' new form and in considerable variety, especially the flowers, many of which do not mature into fruits. Bisexual flowers are rather frequent and. whether bisexual or not, the flowers often have peculiar external markings on the ovary resembling those already noted in the teratological flowers above described. 3. Late flowers, i.e.. abnormal catkins on second-growth wood, have been observed on four of the seven original trees of the new form examined by the writer. The other three trees seldom bear any flowers. Specimens of these abnormal calkins collected from the Leffingwell original tree in 1909 are shown in plate 11. figure 17. Prom the size of the normal young fruit shown at a, the difference in time beiween the normal blooming period and the appearance of these abnormal calkins may be inferred. The flowers on these abnormal catkins are very small and abortive and the leaves, shown at a. b. r. and d. resemble the leaves shown at d, e, and /' in plate 9. figure 14. which were associated with abnormal catkins on indigenous Juglans calif ornica trees. 4. There is no apparenl obstacle to the natural pollination of the late-appearing teratological flowers, as some of them are bisexual and free pollen has been observed in these bisexual flowers. Also a few late staminate catkins have been found. Moreover, there is wide variation in the normal blooming period among individual, indigenous I rei s. The young fruits of different trees have been observed to vary in size from nearly full-grown down to five-eighths of an inch in diameter. It is possible, then, that the normal pollen produced on late trees mighl fertilize alt- normal flowers on early blooming individuals. However, in the lest for apogamic development of normal fruits it would be interesting to test these teratological flowers also. The reasons for not thinking the new variety originated in late teratological flowers are as follows: 1. Sixty-eighl seedlings have been grown from teratological nuts and none have resembled the new form. 2. The new form is known to have been produced by nuts of normal size and shape. From the data at hand, the writer is inclined to consider the second hypothesis as approaching more nearly to the truth than the first, but thus far direct evidence fails to support it. L913] Bdbcock: New Form of Juglans 21 Third Hypothesis The possibility that the original trees of the new form are mutants has been recognized from the first. If it can be shown that the recurrence of this form among seedlings of Juglans cali- fornica trees is due to repeated mutations in certain individuals, this would be good evidence that the original trees were similarly produced. The repetition of mutations is not contrary to experi- ence but rather is characterstic of species known to be in a mutating period. It is generally admitted that plants exhibiting the evidence of being in a mutating period are rare. "Hugo de Vries admet l'hypothese d'une mutabilite periodique et rare" (Blaringhem). Among the plants reported as being in such an abnormal condition, the number of trees is very small. Hence the phenomena connected with the occurrence and recurrence of this new form of walnut gather increased interest as soon as the hypothesis of origin by mutation is considered. The chief obstacle to adequate investigation based on the hypothesis has been the difficulty in locating a single tree of Juglans californica from whose seeds the new form is known to grow. With such a tree located under conditions favorable to experimental work, a careful study of its vegetative and repro- ductive parts can be made and pollination may be controlled. After several years of searching such a tree has now been located. It is a certain tree in the Garden Grove row of black walnuts 1 " from which nuts have been gathered by nurserymen, who have found seedlings of the quercina type during a period of several consecutive years. In 1912 I had the nuts from twenty-one of the trees in this row gathered separately, labeled with numbers corresponding to numbers attached to the trees, and shipped to Berkeley. These were soaked, planted in sand in fiats, properly Labeled and placed under the benches of a glass house, in February, 1913. Two months later, six quercina seedlings were found in the fiats from tree No. 16. There were 275 seedlings in all. Thus slightly over two per cent of the seedlings secured from the 11)12 crop of this tree appear to be mutants. Plate 11, figure 18, shows one of these seedlings and one normal seedling from tree No. 16. Plate 10 Cf. Recurrence of the New Form, j>. 5. 22 University of California Publications in Agricultural Sciences [Vol. 2 12, figure 19, is from a photograph of tree No. 16 taken in April. 1913. About four hundred twigs on this tree likely to bear nuts have been covered so as to insure self-pollination. In this way it is hoped to locate the twig or twigs producing mutants. The data at hand indicate that the new form is not a typical mutation from the seed in the usual meaning of that expression. All but one of the tests of the seeds from the original trees have revealed a partial reversion to the species type in the second generation (counting Disher's original trees as the first genera- tion). Now, one of the generally accepted distinguishing char- acteristics of a mutant is that it breeds true from the first, but here is a remarkably distinct form which does not breed true. Of course, this should not exclude the possibility of origin by muta- tion of an "eversportin"' variety." It is quite possible thai quercina is an eversporting variety and. if so. it may still have originated without an antecedent hybridization, but would then never breed true. 11 However, it does not appear to the writer that the term "eversporting variety" as used by De Vries is applicable to this form. De Vries classes as eversporting varieties such inconstant forms as striped flowers, live-leaved clovers, and polycephalic poppies. But in this new walnut, we have a form which is distinct in most of its characters from the parent form and which breeds true in a portion of its offspring, the remaining portion showing complete resemblance to the parent form. It has been suggested by Professor II. B. Torrev that muta- tions may occur in the gametes of one sex while the gametes of the other sex are normal and, hence, that the new form appeal's among first generation seedlings, but fails to breed true in the second generation. Some doubts as to Professor Torrey's suggestion arise when we consider the results of a pollination experiment which the writer made in 1908. Pollen from Disher's original non-fruiting tree was placed upon pistillate flowers of Tree I, referred to in the discussion of the first hypothesis. All of the eleven seedlings growing from this cross resemble Jittjla/is californica. On the basis of Professor Torrey's suggestion, this would be explained 11 Dr. George H. Shull kindly suggested the explanation on the basis that quercina is an eversporting variety. L913] Babcook: New Form of Juglans 23 by assuming that the pollen grains engaged in fertilization bore -/. call) "ornica characters as a result of segregation in the reduc- tion divisions. But it is as reasonable to explain these results by assuming that the ovnle-borne characters of J . calif ornica are dominant or prepotent over the pollen-borne characters of the new form, supposing that fertilization actually took place. It is obvious that more extensive experiments should be made in an effort to secure reciprocal crosses between the species and the variety. Webber has expressed the following opinion : " Of the various causes of origin it seems to me most reasonable to assume that it is a mutant and the type of mutation to nanate form is similar to Oenothera nanella and the Cupid Sweet Pea. One finds par- allel cases of partial reversion to the parent type among De Vriesian mutations." The reasons for thinking the new form may have originated through mutations in otherwise normal flowers of J. calif ornica may be stated briefly as follows : 1. In the 1911 recurrence of the new form in Garden Grove, it was found by actual examination of the seed bed that all of the aberrant seedlings examined grew from nuts of normal size and shape. 2. The crop of 1912 from a certain tree of Juglans calif or- nica (No. 16 in the Garden Grove row above mentioned) has produced several seedlings of the new form. The nuts from which they grew are of normal size and shape. The possibility of hybridization with any species of oak or other species of walnut is very remote. 3. The large tree standing close to No. 16 is known to have been the source of many of the nuts planted in Tyler's nursery in the years when quercina seedlings appeared in his seedbeds. On this account it was suspected of being the source of the new form. It appears, however, that No. 16 has been the source of some quercina seedlings and that the particular seeds that save rise to it were produced on the branches nearest the tree origin- ally suspected. Transmitted May 9, 1913. PLATE 1 Juglans californica var. querdna Babeock Fig 1. — Disher's original trees in Garden Grove, Cal. The one ;it the left bears regularly, the other produces only staminate or very rarely abortive pistillate flowers. [24] *5! «< *^ o "0 a 03 > o po C/J o < o 03 > CD o o o I 'LATH 2 Juglans California! var. quercina Rabcock Fig. 2. — Leaves and nuts from Disher's original fertile tree. X 1. Fig. '■'>. — Normal leaf (a) fr the Leffmgwell original tire and four leaves from the same tree bearing ascidia. ■ ' .... [26] UNIV. CALIF. PUBL, AGR. SCI. VOL. 2 [BABCOCK] PLATE 2 3 PLATE 3 Juglans calif ornica var. qui rcina Babcock Fig. 4. — Twigs with partially developed fruits from Disher's original fertile tree. X %. .Indians calif ornica var. quercina Babcock Fig. 5. — Twigs with fruits from Disher's original fertile tree and a twig bearing late flowers from the same tree. X %. [28] UNIV. CALIF. PUBL. AGR. SCI. VOL. 2 [BABCOCK] PLATE 3 4 5 PLATE I Juglans ealifornica Wats. Pig. 6. -a. b, r. d, '. nuts sectioned transversely; f, exterior of one Milt. X 1. Juglans ealifornica var. quercina Bal ek Pig. 6.— g, It. i, /. I., nuts sectioned tranversely; ', exterior of one nut. X 1. [30] <** ^s- o-> ^ y^JW ^ C> ^> 3§ai '=-*% : - o> (5 o 1- "0 03 3> Q CO n o CD > CD o o o 7i ^ I- - PLATE 5 Juglans calif omica var. quercina Babcock Fig. 7. — First recurrence of the new form. A two-year-old seedling at Garden Grove in 1907. Fig. 8. — Leaf from another seedling of the same lot as the one shown in fig. 7. X 1. [32] UNIV. CALIF. PUBL. AGR. SCI. VOL. 2 [BABCOCK] PLATE 5 Lr•_.. [36] UNIV. CALIF. PUBL. AGR. SCI. VOL. 2 [BABCOCK] PLATE 7 PLATE 8 Juglans calif ornica Wats. Fig. 12. — a. Ts'ormal pistillate catkin with a developing nut; b. Abnor- mal catkin with pistillate flowers; c. Shows asymmetry of the abnormal flowers; d. Shows depression in surface of ovary; i. A bisexual flower. X 1. [38] o > T) C DO O CO o < O DO 03 n o o 7; PLATE 9 Juglans californica Wats. Fig. 13. — Abnormal fruits — a, b, c, type x; d, e, type y. a', V, c' , d', e'. Abnormal nuts from similar fruits. a", b" , c" , d", c". Similar nuts sectioned transversely and placed in relatively the same position, the basal portions being uppermost in the figure. All X %. Fig. 14. — Abnormal leaves — a, b, c, from second-growth lateral branchlets; d, e, f, associated with abnormal catkins; all collected from wild trees in Brea Canon in 1909. X %. [40J UNIV. CALIF. PUBL. AGR. SCI. VOL. 2 [BABCOCKJ PLATE 9 a d a IB a b jj> *»■ b" 13 d d %■ PLATE 10 Fig. 15. — a, Abnormal leaf on a second-growth lateral branchlet on Juglans calif ornica, Brea Canon, 1909; b, a typical leaf from the Leffing- well quercina tree. X %. Fig. 16. — One of the three original quercina trees at N. B. Pierce's gardens in Santa Ana, 1913. [42] UNIV. CALIF, PUBL, AGR. SCI. VOL. 2 [BABCOCK] PLATE 10 a 15 PLATE 11 Juglans calif ornica var. quercina Babcock Fig. 17.- — Abnormal late catkins from the Leffingwell tree; o, normal catkin bearing one young fruit and with an abnormal catkin growing from its base; b, c, d, abnormal catkins and leaves. X %. Fig. 18. — Two seedlings from Juglans calif ornica tree No. 16 (Garden Grove), March, 1913; a, typical of the species; b, quercina form. X %. [44] UNIV. CALIF. PUBL. AGR, SCI. VOL. 2 [BABCOCK] PLATE II PLATE 12 Juglans califormea Wats. Pig. 19.— Tree No. Hi. Garden Grove, April, L913. I 46 | UNI7. CALIF. PUBL. AGR. SCI. VOL. 2 [BABCOCK] PLATE 12 UNIVERSITY OF CALIFORNIA PUBLICATIONS IN AGRICULTURAL SCIENCES Vol. 2, No. 2, pp. 47-70, pis. 13-19 October 31, 1914 STUDIES IN JUGLANS II FURTHER OBSERVATIONS ON A NEW VARIETY OF JUGLANS CALIFORNICA WATSON AND ON CERTAIN SUPPOSED WALNUT-OAK HYBRIDS BY ERNEST B. BABCOCK In the preceding paper on this subject 1 appeared the history and description of a new form of California Black Walnut, which was named Juglans californica var. quercina, together with a discussion of several hypotheses regarding its nature and origin. The following hypotheses were deemed worthy of serious consid- eration and feasible as bases for further study : (1) Origin through hybridization with the Coast Live Oak, Quercus agri folia Nee. (2) Origin from teratological flowers and fruits of J. cali- fornica. (3) Origin by mutations in apparently normal flowers and fruits of J. californica. In order to test the first hypothesis, that of origin through hybridization between walnut and oak, attempts to hybridize the two species were made in the years 1908 to 1911, inclusive. As a result of the pollination work in 1908, fifty-six seedlings were secured from nuts that developed from pistillate walnut flowers which had been pollinated with oak pollen under bag. Except for considerable variation in size, these trees have not exhibited more differences than would be found among ordinary trees of 48 University of California Publications in Agricultural Sciences [Vol. 2 J. californica growing in the wild. However, since in certain known interspecific hybrids the F t plants all resemble the female parent, it was deemed advisable to grow at least the second generation from all these trees. In 1913 thirty of them bore imts, which were collected and stratified. Also in the spring of 1913 large paper bags were placed on most of these trees enclos- ing flowers of both sexes in order to insure self-pollination of a few flowers on each tree. These bags were shaken vigorously several times during anthesis and upon examination of several flower clusters the stigmas were found to be well covered with pollen. Very few nuts developed under bag, but this may not have been due to self -sterility on the part of the individual trees. A period of very warm and humid weather caused most of our walnut flowers that were developing under bag, both in Southern California and at Berkeley, to drop. From the self-pollinated nuts secured on these supposed hybrids 75 seedlings from 24 different trees are growing. These, together with 2001 seedlings from naturally pollinated nuts representing 30 different trees. give a total of 2076 second generation seedlings that have been examined. Among them all not a single individual has been found that resembles the new variety. Neither is there any indi- cation of oak characters in these F. 2 seedlings. Therefore as yet we have no evidence that the F t trees are true hybrids. A suffi- cient number of F 2 plants will be retained to grow the F :! gener- ation from each F t tree. Now the question arises as to the nature of the supposed F, hybrids. I consider the conditions under which the work of pollination was performed in 1908 to have been practically ideal. The pistillate flowers were bagged long before any walnut trees were shedding pollen. On one of the two trees I am certain that no pollen was being shed even at the time I pollinated. This tree was fairly well isolated from other walnuts, some of which were shedding pollen at that time. Moreover, no nuts developed in the check bags on this tree. Now if natural polli- nation was prevented both before and after the application of oak pollen, why are not these trees hybrids ? Are they the result of abnormal embryogeny due perhaps to stimulation by the oak pollen but without the occurrence of fertilization ? It should 1914] Babeock: New Variety of Juglans 49 be remembered that they are the progeny of only two different trees and. if they are the result of asexual reproduction, it would be reasonable to expect among the offspring of either parent marked uniformity in size, leaf characters, time of putting forth and shedding leaves, time of flowering and flower characters. But the variations in the above mentioned characteristics are so great as to suggest heterogeneous parentage within the species rather than asexual reproduction from one or two parent trees. Plate 13. figure 1, and plate 14, figure 2, show a typical leaf and a cluster of partly developed fruits from each of six of these F l seedlings. With these facts in mind, let us consider briefly the various processes of abnormal embryogeny that may have given rise to these variable seedlings. (1) The new sporophyte may have developed from the megaspore mother-cell, in which case its cells would possess the diploid number of chromosomes charac- teristic of Juglans calif ornica, provided that the parent tree was typical of the species. (2) It may have arisen from the mega- spore, from the egg nucleus, or from one of the other embryo sac nuclei, without fertilization, in which case its cells would contain the haploid number of chromosomes. (3) It may have arisen adventitiously from sporophytic tissue, in which case its cells would contain the diploid number of chromosomes. It is obvious that in either the first or third instances we should ex- pect much uniformity among the progeny and close resemblance to the parent trees. Hence, in view of the wide variation men- tioned above, it is reasonable to assume the occurrence of one or more of the three phases of parthenogenesis included in the second of the three cases above defined. Finally, this hypothesis of origin by hybridization with oak is practically annihilated by the discovery previously reported that in a row of twenty-one California Black Walnuts, growing in Garden Grove, Orange County, California, a single tree has been found which produces the new variety year after year. I know of no oaks in this region, but even if oaks were abundant and close at hand, the fact that quercina seedlings come from only one tree would certainly indicate some other cause than oak pollination. 50 University of California Publications in Agricultural Sciences [Vol. 2 As for the second hypothesis, that the new variety originated from teratological flowers of J. calif omica, this seemed very unlikely after testing sixty-eight abnormal nuts from several different trees and failing to secure it, and especially after dis- covering that the new form does originate from nuts of normal appearance. At the same time, the occurrence of teratological leaves, flowers, and fruits in this species would certainly indicate an unstable condition, which finds its most frecpient expression in these abnormal features of the somatoplasm and which may occasionally result in such segregations in certain cell divisions preceding or accompanying gametogenesis as would result in what we call mutation. As a matter of fact the tree of J. cali- fomica, mentioned above, which has been under observation about two years (tree no. 16 mentioned at the close of the preceding paper), produced in 1913 a few clusters of these abnormal late nuts. These were found while gathering the normal nuts from this tree. Each cluster (produced on a single catkin) was gath- ered separately and given a number. A tree label bearing the same number was attached to the twig that bore that particular cluster. From each of three of these clusters of teratological fruits two or more quercina seedlings have appeared. Plate 15, figure 3, shows the seedlings grown from abnormal cluster number B4 which was borne on a late-appearing catkin which sprang from the base of a normal pistillate catkin on twig number 151. In this case there are four typical calif omica plants {a, b, c, d), four quercina plants (e, f, g, h), and two seeds that did not germinate. It may be questioned whether seedling h could be properly classified at such an early stage. Suffice it to say that, even at the time when the stem is just pushing through the soil, the appearance of quercina is quite distinct from that of cali- f omica. Seen from above the former has the shape of a rosette while the latter appears as a cone. This is due to the decided difference in the apices of the leaves (cf. b and e). Seedling h presents a fine exhibition of geotropism due to the position in which the seed happened to be planted. The extreme abnormality of some of the nuts of this cluster is strikingly shown in a, c, e and /. In each of these seeds one cotyledon was confined within a sector equal to about one-third the volume of the nut. It is 1914] Babcock: New Variety of Juglans 51 obvious that there is no relation between degree of abnormality of the nut and the appearance of the quercina form. However, the fact that quercina seedlings have been secured from terato- logical fruits might seem to indicate that in this fact of tera- tology one finds the basis for a complete solution of the problem in hand. But these are the first quercina seedlings I have raised from abnormal nuts, and of the twenty-eight seedlings secured only eleven are quercina in character. On the other hand, all the young quercina seedlings examined in the past have sprung from apparently normal nuts. This shows that something more than the mere fact of teratology must be found to explain the origin of the new variety. Regarding the third hypothesis the evidence now at hand is definite and sufficient. Over three hundred clusters of normal nuts were gathered separately from tree No. 16 in 1913, fol- lowing the method described above. The number of nuts per cluster varied from one to five. The nuts of each cluster were stratified in a pot bearing the same label as the twig from which the cluster came. Later the pots were transferred to a cool green- house where they were kept until the plants were several inches high. One quercina seedling was found in each of 42 pots. Plate 16, figure 4, shows the seedlings from the two nuts in cluster 35 ; plate 17, figure 5, shows the seedlings from cluster 97 ; and plate 18. figure 6, the seedlings from cluster 196. Besides the 600 seedlings grown from marked clusters of nuts, about 1000 addi- tional seedlings were raised from this tree. Of the total num- ber of seedlings grown approximately 5 per cent were quercina. The fact that among the normal fruits of this particular tree only one nut in a cluster produces the quercina form, at once suggests a possible relation between location in the cluster and production of quercina seedlings. Observations and experiments are being conducted in order to determine whether there is any definite position in the cluster or other morphological feature that is associated with origin of the new form. The nuts gathered from this tree in 1913 may not have been self-pollinated. About four hundred branchlets including staminate and pistillate flowers were bagged in early spring, but the warm moist weather mentioned above caused all the pistillate 52 University of California Publications in Agricultural Sciences [Vol. 2 flowers to drop. Thus my only recourse was to label and bag clusters of developing nuts. About 350 clusters were so treated. It is very likely that many of the flowers were cross-fertilized with pollen from neighboring trees. However, if the pollen of either no. 16 or its near-by neighbors had been the source of the new form, quercina seedlings should have been obtained among the progeny of other trees besides no. 16 in the original seed test made in 1912. Since only the one tree produces the new form I have disregarded pollination for the present, studying only the pistillate flowers and testing the seeds. Cytological investigation will possibly reveal the true nature of the new form and perhaps explain its origin. Meanwhile the speculations concerning the nature of the supposed walnut-oak hybrids are suggestive in this connection. Through failure of pollination or fertilization parthenogenesis may take place. That polyembryony occurs both in the new form and the old is proved by the discovery of the specimens shown in plate 19, figures 7 and 8. In each case the two embryos were complete, each caulicle being attached to its own pair of cotyledons. Again, it is possible that at some stage in flower development abnormal mitosis oc- curs. Should this happen in a very early stage in the develop- ment of the flower, sufficient abnormal somatic tissue would be produced to make cytological investigation comparatively easy. It is hoped that by determining the morphological location of the nuts that produce quercina seedliims the cytological study of very young flower clustei's will lie somewhat simplified. The present tendency 2 to refer to hybridization as the basis of all variation calls for a reference to my previous paper 1 where I discussed the hypothesis of origin through hybridization either with oak or with any form of Jiii/laiix and allied genera, and showed that such a hypothesis is untenable. The results reported in this paper bear out that conclusion. Further, as opposed to the proposition of assuming hybridization as necessary for the occurrence of mutation, we have the reeenl conclusion of Gates 3 "that mutation and hybridization are separate phenomena, and that the cause of some at least of the mutations in (Enolhera is independent of the combination of hybrid characters." 1914] Babcoclc: New Variety of J 'uglans 53 The origin of quercina is similar to those transmutations in Lycopersicum, 4 Gossypium, 5 Nicotiana, and Oenothera, 3 which have been described as aggregate mutations as distinguished from loss or addition of single characters as, for example, in Helianthus 7 and Drosophila* In regard to the tobacco mutation above referred to, the authors 6 assume ' ' that mutation must have taken place after fertilization, i. e., after the union of the male and female reproductive cells." Castle 9 suggests it is equally probable that the mutation occurred in an egg cell which then developed without fertilization since "parthenogenesis is known to occur in tobacco and mutation in a growing or immature germ cell seems inherently more probable than in a fully formed and fertilized one." This discussion is pertinent to both phases of the problem herein set forth, viz., first, an explanation of the variation found in the F x oak-pollinated walnut seedlings ; second, the cytological time of the mutations that produce quercina seedlings. Concerning the first question, parthenogenesis or, more specif- ically, apogamy, assuming the occurrence of reduction in chromo- some number, might explain the variation in the F x oak-pollinated seedlings whereas such an extent of variation is too great to permit the assumption that embryos developed from the spore mother cell or other sporophytic tissue without also assuming irregularities in chromosome behavior. Reduction is not assumed 10 in the classical cases of parthenogenesis in angiosperms (Anten- naria, Taraxacum, Hieracium and Alchemilla) , but in Thalicirum purpurascens Overton 11 finds "the development and germination of the megaspore is that usually found among angiosperms." He reports no observations on chromosome behavior but shows that "parthenogenesis is becoming fixed in Thalictrum." In such an instance it seems reasonable to assume that the omission of reduc- tion has become established also. But in Juglans the existence of adaptation for wind pollination would indicate that pollination and fertilization are usually essential for seed production and there is experimental proof 1 of this also. Moreover, the extensive researches of Nawaschin and Finn 12 on the cytology of fertiliza- tion in Juglans, although lacking the treatment of chromosome number and behavior, demonstrate that the process of fertiliza- 54 University of California Publications in Agricultural Sciences [Vol. 2 tion is typically chalazogamous, there being a well-developed bi- nucleate cell which reaches the embryo sac. Hence the occurrence of reduction may be assumed, in which case embryos must arise through apogamy in the absence of fertilization. But with the haploid number of chromosomes in the somatic tissue of these seedlings, as will be shown below, one might expect even greater deviations from the parent forms than actually occur. Thus, if apogamy and reduction be assumed, it seems necessary also to assume 13 "the subsequent arrest of the homotypic mitosis (in the cell destined to differentiate the embryo sac) before the division of the nucleus has taken place, resulting in the pro- duction of a functional germ cell with a chromosome number double that of its reduced number. ' ' Regarding the other question, the mutations that produce quercina seedlings evidently occur only in pistillate calif ornica flowers, thus producing seedlings that will not breed true. This indicates that the cytological time of mutation is previous to fertilization. From her cytological studies of Oenothera Lutz 14 has con- cluded that "all individuals of a given type of vegetative char- acter invariably have identical somatic chromosome numbers regardless of the diversity of origin of the individuals in ques- tion," and, further, that "All individuals . . . having a chro- mosome number much in excess of that in O. Lamarckiana dis- played certain characters strongly suggesting those of 0. gigas, chiefly noted in the stoutness of all parts." This suggests that in a form like quercina, which is reduced in all vegetative char- acters, we should expect to find the somatic number of chromo- somes to be less than the number typical of the species. The occurrence of such a mutation in Juglans is of especial significance because of the phylogenetic relationships ascribed to these chalazogamous forms. According to Nawaschin and Finn, 12 the preservation of the male cytoplasm in the species of Juglans indicates an old tendency inherited from gymnosperm ancestors and furnishes further important proof of the great age of these forms, which stand on the threshold of the angiosperm world. Berry 15 mentions the occurrence of seven species of Juglans in the upper Cretaceous deposits, twenty-five in the 1914] Babcock: New Variety of Juglans 55 Eocene, several in the Oligocene, upwards of twenty in the Miocene, and about twenty-five in the Pliocene, several of the latter being very close to, if not identical with, existing species. These are what remain in the very imperfect geological record. Doubtless many others existed. Now that aggregate mutation is known to have occurred once in such a group, it is reasonable to assume the occurrence of such major discontinuous variations as one of the processes by which new species have been produced. A recent effort 1 " to harmonize the older theories of evolution through continuous variation and the modern conception of alternative inheritance assumes that "unit characters" are really composite in nature, but the paleobotanist will not necessarily accept the hypothesis that all new species were built up from such minute discontinuous variations that the effect is one of continuous though gradual change. It is just as reasonable to assume that the fossil species sprang into existence in the same sudden manner as that by which quercina made its appearance. On account of the persistent assumption by some 17 that quercina is a natural hybrid between walnut and oak and that the progeny of the original quercina trees are composed of walnut and oak seedlings in Mendelian proportions, it should be noted that there is wide variation among the fruiting specimens of the new form as to the proportion of californica and quercina seed- lings they produce. I have tested seeds from three of the original trees distributed by Disher 1 and find they differ very widely in this respect. They certainly produce no oaks and there is no basis for assuming a Mendelian ratio among the progeny. 56 University of California Publications in Agricultural Sciences [Vol. 2 SUMMARY 1. The belief, which has been held by many, that the new form of walnut, Juglans californica var. quercina, originated through hybridization between walnut and oak is without founda- tion in fact. There is no evidence that hybridization with other species of walnut or cross-pollination with other trees of the same species causes the appearance of the new form. 2. Although abnormal, late appearing flowers and fruits oc- cur rather frequently on the California Black Walnut, only one tree has been found to produce abnormal nuts which give rise to quercina seedlings, and more than half of the abnormal nuts from this particular tree produce typical californica plants. Certainly teratology cannot be considered the cause of the origin of the new variety. 3. The evidence that seedlings of the new variety come from certain apparently normal nuts is conclusive, since a tree has been located that annually produces a small percentage of quer- cina seedlings. Evidently these aberrant individuals are the re- sult of internal changes that take place during the growth of the flowers previous to fertilization. It is possible that evidence of these changes will be discovered by microscopical study of material from tree no. 16 and that further breeding experiments will help to explain them. Meanwhile, whether the true nature of these changes be revealed or not, we know that embryos are produced independently of the influence of self- or cross- pollination, which are capable of developing into individuals pos- sessing characters strikingly different from those of the parent and capable of transmitting those same characters to at least a portion of their progeny. Such occurrences of discontinuous variation are generally recognized as mutations, and the quercina walnut is similar to certain mutations in tomato, cotton, tobacco and evening primrose, which have been designated as aggregate mutation. Transmitted June 15, 1914. 1914] Babcock: New Variety of Juglans 57 PUBLICATIONS REFERKED TO IN THIS PAPER 1 Babcock, E. B., Study of a New Form of Juglans calif omica Watson, Univ. Calif. Pub). Agric. Sciences, vol. 2, no. 1, 1913. 2 Bateson, Wm., Address of the President of the British Association for the Advancement of Science at Melbourne, Science, n. s., vol. 40, no. 1026, p. 296, 1914. s Gates, R. R., Breeding experiments which show that hybridization and mutation are independent phenomena, Zeitschr. f. Ind. Abstamm.-u Vererbungslehre, bd. 11, hft. 4, 1914. 4 White, Chas. A., The Mutations of Lycopersicum, Pop. Sci. Mo., vol. 67, no. 2, p. 151, 1905. s White, Chas. A., Aggregate Mutations in Gossypium, Science, n. s., vol. 27, no. 683, 1913; and Cook, O. F., Mutative Reversions in Cotton, Circ. no. 53, Bur. Plant Industry, U. S. Dept. Agric, 1910. 6 Hayes, H. K., and Beinhart, E. G., Mutation in Tobacco, Science, n. s., vol. 39, no. 992, p. 34, 1914. 7 Cockerell, T. D. A., The Red Sunflower, Pop. Sci. Mo., vol. 80, no. 4, p. 373, 1912; A Wine-Eed Sunflower, Science, n. s., vol. 38, no. 974, p. 312, 1913; Suppression and Loss of Characters in Sunflowers, ibid., n. s., vol. 40, no. 1025, p. 283, 1914. s Morgan, T. H., Heredity and Sex, 1913. Castle, W. E., The Cytological Time of Mutation in Tobacco, Science, n. s., vol. 39, no. 995, p. 140, 1914. i° Strasburger, Noll, Schenck and Karston, A Text-Book of Botany, p. 93, 1908. n Overton, Jas. B., Parthenogenesis in Thalictrum purpurascens, Bot. Gaz., vol. 33, p. 363, 1902. i 2 Nawaschin, S., and Finn, V., Zur Entwickelungsgeschichte der Chala- zogamen, Juglans regia and J. nigra, in Mem. Acad. Imp. Sci. St. Peters- bourg, 31, pp. 1-59, 1913. Also noted in Bot. Gaz., vol. 57, no. 2, p. 162, 1914. is This is the 5th condition of Gregoire's hypothesis for explaining chromosome numbers in (Enothera as set forth by Lutz (cf. 14, p. 426). i* Lutz, Anne M., Triploid Mutants in Oenothera, in Biolog. Central., bd. 32, no. 7, 1912. 15 Berry, Edw. W., Notes on the Geological History of the Walnuts and Hickories, Plant World, vol. 15, no. 10, 1912. 10 Collins, G. N., Nature of Mendelian Units, Jour. Heredity, vol. 5, no. 10, 1914. 17 Pierce, N. B., A New Walnut, Science, n. s., vol. 37, no. 955, p. 613, 1913. PLATE 13 Juglans californica X Quercus agrifolia Fig. 1. — a, leaf and fruits from F seedling No. 13a. Fig. 1. — b, leaf and fruits from F seedling No. 13b. Fig. 1. — e, leaf and fruits from F seedling No. 13c. The nuts from which these three F seedlings grew were borne in the same cluster on J. californica tree I of 1908 hybridization experiments. X %. [58] > "0 CI CD > CO o < o CD > CO n o n PLATE 14 Juglans californica X Quercus agrifolia Fig. 2 — a, leaf and fruits from F! seedling No. 23a. Fig. 2 — b, leaf and fruits from F, seedling No. 23b. Fig. 2 — c, leaf and fruits from F, seedling No. 23c. The nuts from which these three F, seedlings grew were borne in the same cluster on J. californica tree II of 1908 hybridization experiments. X y 3 . [60] < o ~0 o o o PLATE 15 Juglans californica Watson J. californica var. quercina Babcock Fig. 3. — a, b, c, cl, typical californica seedlings, — e, f, g, h, typical quercina seedlings. All from the same cluster of abnormal fruits from tree No. 16 in Garden Grove, Calif. X %. [62] o > ~0 CD 3> O XI GO o < o CO 3> CD o o o PLATE 16 Juglans calif ornica Watson J. calif ornica v;ir. qu. I 68 | o > C CD 3> c< XI C/) o < o CD O o o ~V I - > PLATE 19 Juglans californica Watson Fig. 7. — Specimen of polyembryony from cluster No. 208 on ,/. cali- fornica tree No. 16 in Garden Grove, Calif. X '_•• Juglans californica var. quercina Babcock Fig. 8. — Specimen of polyembryony from a cluster of abnormal nuts on J. californica tree No. 16 in Garden Grove, Calif. X ' •_>. [70] UNIV. CALIF. PUBL. AGR. SCI, VOL. 2 [BABCOCK] PLATE 19 UNIVERSITY OF CALIFORNIA PUBLICATIONS IN AGRICULTURAL SCIENCES Vol. 2, No. 3, pp. 71-80, plates 20-21 September 20, 1916 STUDIES IN JUGLANS, III (1) FURTHER EVIDENCE THAT THE OAK-LIKE WALNUT ORIGINATES BY MUTATION BY EENEST B. BABCOCK In previous studies 1 the conclusion was reached that the oak- like walnut, Juglans califomica var. quercina Babcock, was first produced as a result of germinal variation in a tree of the southern California black walnut, J. califomica Wats., and that the several recurrences of this peculiar walnut resulted from repetitions of the same mutation. It is the purpose of this paper to report further evidence showing that this conclusion was correct at least as regards the origin of the first quercina individual, although this same evidence may lead to a different explanation of some of the recurrences of this form. This evidence consists of the results from hybridizing quercina and califomica. In 1908 pollen from the original, fertile, type individual of quercina was applied to several guarded pistillate flowers on a typical califomica tree. The eleven P x seedlings secured are growing on the campus of the University of Cali- fornia. They are all normal califomica trees. The only evidence of heterosis thus far observed is length of staminate catkins, which is intermediate between the two parents. In 1915 several of these F x trees were self-pollinated, or interpollinated, and thirty-six seeds were secured. Up to the present eighteen of these have germinated, producing twelve califomica and six quercina 1 Babcock, E. B., Studies in Juglans I and IT, Univ. Calif. Publ. Agric. Sciences, vol. 2, no. 1, 1913; no. 2, 1914. 72 University of California Publications in Agricutlurdl Sciences [Vol.2 seedlings, which is a ratio of 2.67 to 1.33, a deviation from the theoretical monohybrid ratio which might be expected to occur under the laws of chance in 42 per cent of monohybrid popu- lations. While positive statements will not be made until the F, trees have been tested on a more extensive scale, yet a valid :{ to 1 ratio even with so few F 2 individuals is certainly an indication that the genetic relationship between califomiea and guercina is a difference in a single factor of the same Mendelian reaction system. The idea that a single factor-difference may so affect the entire chromosome system that the individual is altered more or less in each somatic detail is now generally recognized, yet the direct evidence on which this conception is based is found in a limited number of cases. Morgan 2 refers to the mutant strains of Drosophila a>»p< lopltila called "club" and "rudimentary", in which the factor for a certain wing character also conditions the development of certain other morphological and physiological characters. But, thus far in Drosophila, no single factor has been found thai visibly affects every external feature of the organism. The well-known dwarf or cupid sweet pea is a strik- ing example of the manifold effects of a single factor. This variety differs from the ordinary climbing form of Laflnjnts odoratus not only in its extremely dwarf stature but also in color of foliage, length of internodes, size and arrangement of flowers, time of anthesis, fertility and viability. Yet it is certain that the variety differs from the species in a single genetic factor. Probably this is as striking a case as has been reported previously, yet in such a conspicuous character as leaf-shape the dwarf variety very closely resembles the species type. Now the oak-like walnut differs from the species type in every gross external feature — shape of leaves, color of foliage, color of bark, habit of growth, structure of inflorescence, struc- ture of flowers, size and structure of fruits, as well as in fertility and viability. Therefore, if tests that are now being made confirm our inference that quercina differs from the species in a single genetic factor, it will be a most striking example of the manifold effects of a single genetic factor. Further- 2 Morgan, T. H., Mechanism of Mendelian Heredity, p. 209ff., 1915. 1916] Babcock: Studies in Juglans, III 73 more, the demonstration of such a genetic relation between quercina and californica must be accepted as ample proof that the first quercina individual, at least, originated by mutation. Since the mutant factor is recessive to its normal allelomorph, it is highly probable that californica trees which are known to produce quercina seedlings are not to be considered as mutating individuals. Such a tree is ./. californica, "Garden Grove No. 16", which was referred to in the preceding number of this series. This tree may he either the result of a mutation in on* 1 of the gametes that produced it. in which case it would, of course, lie heterozygous, or it may he a heterozygote produced by hybridization between quercina and californica. The same is time of the other two californica trees known to have produced quercina. But the genetic relation between the variety and the species shows that the first production of quercina at least must have been caused by a mutation in one genetic factor and that this change occurred in all probability in a germ cell of the grandparent of the first quercina tree. A question arises as to the interpretation of the test of the particular californica tree. Garden Grove, No. 16, reported in the preceding paper. In 1913 the unguarded seed from this tree produced about 5 per cent of quercina seedlings. The fact that the seed was not self-pollinated under control and so may have been crossed with nearby trees may serve as an explanation of this result. However, another explanation is indicated. This tree is very late in developing its pistillate flowers, the stigmas becoming conspicuous after most other trees have shed their pollen and after this tree has shed much of its own pollen. Xow if only a portion of its pistillate flowers are self-pollinated the remainder probably develop apogamously. That would account for fewer qut rcina seedlings than would be expected from self-pollinated seeds in case this tree is heterozygous for the quercina factor. For only one-fourth of the self-pollinated nuts would produce quercina seedlings, whereas all apogamous seeds would produce californica individuals. Presumably all such apogamic plants would contain the diploid number of chromo- somes. Hence they would be heterozygous like their parent and for this reason they would be californica in type. 74 University of California Publications in Agricutlural Sciences [Vol. 2 There is also the remote possibility that unfertilized pistillate flowers might develop parthenogenetieally, the egg-nucleus, containing the haploid number of chromosomes, developing spontaneously into an embryo. In order that such seeds should produce only calif ornica individuals it would be necessary that a single dose of the quercina factor could not condition quercina development even in plants having the reduced number of chromosomes. This is at variance with the concept that Men- delian reaction systems depend upon proportional chemical relations such that one dose of a recessive factor would play the same role in an individual having a haploid system as two doses of the same factor play in an individual having a diploid system. Therefore, the former is the more reasonable inter- pretation. (2) A PARALLEL MUTATION IN JUGLA.XS IlfXHSII (JEPSON) SARGENT In November. 1914, through the courtesy of Farm Adviser F. F. Lyons, my attention was called to a nursery at Modesto, California, where there were several thousand one-year-old seed- ling walnuts. Here and there among the typical black walnuts I found a number of plants (fifty or more) that closely resembled ./. californica var. quercina except that they were taller than qui rcina seedlings of the same age and the leaves appeared somewhat larger. Through the kindness of the owner, George F. Covell, seven of these seedlings are now growing on the campus of the University of California. These seven and several that were examined at the nursery were found to have come from typical nuts of J. Jiiiirfsii, the northern California black walnut. The trees that produced the nuts which Covell planted in bis nursery are also typical of J. hindsii, but as they had been grafted to commercial varieties I was unable to secure seeds from them. However these grafted trees are seedlings from four large north- ern black trees growing near Lodi, California. Several hundred nuts from each of these trees have been germinated and only seedlings typical of J. hindsii have been secured. If any one of these trees is repeating the mutation there is no evidence of it in 1916] Babcoch: Studies in Juglans, III 75 the immediate progeny. This is what would be expected if the new variety has the same genetic relation to J. hindsii as quercina has to J. calif or nica. In an earlier paper 3 I proposed to designate quercina as J. californica mut. quercina and the quercina-like form of hindsii as J. californica var. hindsii mut. quercina. However, the recog- nition of hindsii as a species 4 simplifies the problem and makes it desirable to describe the new mutant from hindsii as a variety of that species and to retain quercina as a variety of californica. The following description is based upon material gathered from several of the seedlings in Covell's nursery in 1914. The seven seedlings growing on the campus of the University of California are cited as cotypes. It should be noted that the variety name has been chosen for the express purpose of emphasizing the fact that the new variety resembles quercina in leaf characters. New Variety Juglans hindsii var. quercinifolia Babcock Tree. Bark and leaves strongly walnut-scented. Pits in plates. Twigs, bud scales, and young leaves granular pubescent. Buds few-scaled axillary or superposed. Leaves 1 to 31/2 inches long, alternate, exstipulate, mostly compound with three leaflets ; terminal leaflet 1% to 2 times as long as lateral leaflets and ranging from % to 2 1 / i inches in length, in form ovate or elliptical, obtuse or truncate at the apex, margin irregularly crenate or serrate ; lateral leaflets mostly opposite and sessile, sometimes one or both lacking, occasionally one or two extra ones present ; petiole equal to or shorter than lateral leaflets ; very rarely with unifoliolate leaves. (Cf. plate 20, fig. 1.) Nursery of George F. Covell, Modesto, Cal., Nov., 1914, Univ. of Calif. Herb. no. 189541. Cotypes on campus of the Univ. of Calif. (Cf. Div. of Genetics nos. 755a to 755gr.) Plate 20, fig. 1 shows a specimen of quercinifolia which was supplied by Covell in 1915. A typical quercina seedling is shown in plate 20, fig. 3. The relative size of these two seedlings is of no significance as they were not of the same age. In order to s Babcock, E. B., Walnut Mutant Investigations, Proc. Nat. Acad., vol. 1, p. 535, Oct., 1915. •ijepson, W. L., in Smith, K. E., Bniv. Calif. Agr. Exp. Sta. Bull. 203, p. 27 (1909"). Juglans californica Wats. var. hindsii Jepson in Bull. So. Cal. Acad. Sci., vol. 7, p. 23 (1908). " 7(> University of California Publication* in Agricutlural Sciences [Vol. 2 emphasize the fact that hindsii is distinct from calif arnica even in the young seedling stage, a typical seedling of calif ornica and one of hindsii are shown in plate 21. figures 5 and G. The differences between the nuts of the two species are clearly shown. The mature trees of the two species are also strikingly distinct, calif ornica being always low and shrub-like in habit while hindsii is tall and arboreous in form. These parallel mutations in two distinct species may appear as degressive rather than regressive variations or. in other words, as cases of reversion to a common ancestral form. If both quer- cina and quercinifolia resulted from a change in one genetic factor, as seems likely, and both represent a common ancestral form, then it would follow that both calif arnica and hindsii sprang fully formed from their common ancestor by mutation. Yet both quercina and quercinifolia show reduction in morpho- logical characters and quercina individuals exhibit low fertility. These symptoms would indicate that the mutation is regressive rather than degressive. However, the fact that any walnut varie- ties originate by mutation is of significance for the student of evolution, because the Juglandacae are generally considered as one of the older and more stable groups of angiosperms. They are not supposed to be undergoing changes similar to changes that give rise to new types in the younger, less stable groups. That the origin of these two unique walnuts, or of quercina at least, cannot be explained on the basis of hybridization is now fully proved. Evidently quercina sprang from calif ornica rather than calif ornica from quercina, and il arose as a result of a change in a single genetic factor, i.e.. of mutation in the strict sense. Transmitted May 31, 1916. PLATE 20 Juglans hindsii var. quercinifolia Babcock Fig. 1. Specimen from Covell's nursery, 1915. Note typical hindsii nut, the husk removed to show the smooth surface, x A. Fig. 2. Halves of typical hindsii nut. x i. J. calif ornica var. qu&rcina Babcock Fig. 3. Seedling of J. calif ornica, "Garden Grove No. 16." Note typical calif ornica nut. x 1. Fig. 4. Halves of typical californica nut. x J. [7*1 UNIV. CALIF. PUBL. AGRI. SCIENCES, VOL. 2 [BABCOCK] PLATE 20 C\J en oo co m CO cJ PLATE 21 Juglans calif ornica Watson Fig. 5. Typical seedling, x J. J. hindsii (Jepson) Sargent Fig. 6. Typical seedling, x J. [SO] UNIV. CALIF. PUBL. AGR. SCIENCES, VOL. 2 [BABCOCK] PLATE 21 6