Journal of Heredity 48: 11-20 (1957)

* Contribution No. 2772 of the Maryland Agricultural Experiment Station, Department of Horticulture. Scientific Article No. A-597.

AN earlier paper reported the finding of widely differing chromosome numbers in individual plants in the progenies of Boysen and of Young blackberries crossed with colchicine-treated Eldorado1. Two characteristics of these plants implied the action of some mechanism resulting in mitotic instability: 1) the absence of appreciable numbers of chromosome number determinations which could have been derived through polysomaty; and 2) the finding of complete and wide ranges of aneuploid chromosome numbers in single individuals.

Subsequently, additional unstable plants of Rubus have been found in the progenies of many different crosses. These plants and their asexually propagated clones have been observed and studied cytologically to determine the mechanisms causing mitotic instability.

Materials and Methods

All the unstable Rubus plants discussed in this paper were grown from seeds obtained by carefully controlled crosses made in the greenhouse. Leaf-bud cuttings of the mitotically unstable plants were propagated in coarse sand under constant mist in an outdoor propagating frame. Root-cuttings were made in the usual manner of commercial practice. All seedlings and clones were grown in individual pots in the greenhouse.

The cytological methods for root-tip determinations were the same as used by Britton and Hull1. Also, the axillary leaf-bud squashes were made in the same manner. In addition, many methods were tried in order to show the mitotic spindles clearly as well as the chromosomes. While none were completely satisfactory, the best results were obtained by fixing root-tips in CRAF for 35 minutes, rinsing in running water for 5-10 minutes, and storing in 70 percent ethyl alcohol until such time as the root-tips were required. The root-tips were then hydrolized for 20 minutes in normal hydrochloric acid at 60° C., stained in leuco-basic fuchsin (Feulgen) for two hours, and squashed in iron aceto-carmine.


Crosses which have produced one or more mitotically unstable plants among their respective progenies are presented in Table I. It has not always been possible to recognize such unstable individual plants on the basis of their phenotypes. However, many have a pronounced "checking" on their leaves (Figures 5B and 6). These "checks" or mosaic areas are of a lighter green color and are thinner than the rest of the leaf, but rarely involve as much as an entire leaflet. While all plants exhibiting this mosaicism have been found to have various chromosome numbers in their roots, a few mitotically unstable plants have been found with no evidence of "checking" in their leaves.

In one population of 66 seedlings of Eldorado (4x), (2n=28) selfed, five of the young plants were determined to be 2n=42 by means of axillary leaf-bud squashes. These 6x plants presumably resulted from fertilization of unreduced eggs by reduced pollen grains—an event not uncommon in Rubus2. In contrast to the regular-appearing leaves of 4x Eldorado selfed seedlings (Figure 5), the 6x seedlings exhibited, in addition to their symptoms of higher ploidy5 a noticeable degree of "checking" or mosaicism in their leaves (Figure 5B) Cytological examination of their root-tips indicated that many cells had 42 chromosomes and others had fewer—some markedly fewer (Table II). No evidence of mitotic instability has been found in 4x Eldorado plants or 4x selfed seedlings of Eldorado.

Figure 5
A—Leaf of seedling (4x) of Eldorado selfed to compare with B. B—Leaf of seedling (6x) of selfed Eldorado. Note the change in shape, serration, angle of leaf veins, and puckering. The arrow points toward a thinner, lighter area of mosaicism. Two other such areas are present in the lower right leaflet.

As in the case of the 6x seedlings of selfed Eldorado, most unstable plants of other Rubus seedling populations have not been notably weak, dwarf, or otherwise abnormal in appearance. To the contrary, many were among the most vigorous of their respective populations. However, a few dwarfs have been found with unexpectedly high chromosome numbers which appeared most frequently to have resulted from fertilization of unreduced eggs. Many of these produced root suckers with lower chromosome numbers which had markedly increased vigor.

Leaf-bud cuttings of the 28 original unstable plants all grew into plants having first season shoots which appeared to be identical phenotypically to those of the corresponding original mother plants. Chromosome number determinations made from root-tips of these clones indicated that they were still mitotically unstable.

Leaf-bud cuttings were made from these plants that had been grown from leaf-bud cuttings, and again the phenotypic characteristics of the original shoot of the mother plant were maintained. However, cytological studies of their root-tips showed that the mitotic instability was still present. Chromosome number determinations from root-tips made when these latter plants were only a few weeks old revealed many high numbers as well as lower numbers. Later in the season, it was often possible to find only low chromosome numbers (e.g. 20-35). At this later time, axillary leaf-bud squashes were made which showed many cells with the high chromosome number of the original shoot of the mother plant.

On the other hand, root-cuttings from single unstable plants have produced plants exhibiting a wide range of phenotypes quite unlike the mother plant or its parents. In the case of one of the plants previously reported (Plant No. 1, Table III, Britton and Hull1), 27 plants were produced from root-cuttings. Two were weak and died shortly after sprouting and the remaining 25 could be classified into four groups of quite different phenotypes with minor variations within each group. None were at all similar to the phenotype of the original shoot of the mother plant.

Figure 6
Leaf from a mitotically unstable hybrid (Boysen X (8x) Eldorado). Arrows point to some of the more striking areas of mosaicism. The whole area of the leaflet below the midrib and to the left of the arrow at lower right, is of similar appearance to the smaller areas of "checking" on other leaflets,

An example of three canes from the roots of another unstable plant is shown in Figure 7. The leaves on primocane B are thick, irregularly toothed, puckered, and divided into 3-5 leaflets haphazardly. Those on primocane A are extremely uniform in appearance. All are thin, regularly toothed, smooth, and divided neatly into three leaflets. Primocane C has leaves which are markedly "cutleaf" or incised, in a way unlike the mother plant or either of its parents. Further, the cane is more spindly, wiry, and weak than that of either A or B.

Twenty leaf-bud cuttings were made from primocanes A, B, and C (Figure 7), and the plants which grew from these leaf-bud cuttings had the phenotypes of the canes from which they were propagated. Cytological studies of root-tip squashes of the five plants from these leaf-bud cuttings of primocane B showed that in each plant many cells had 56 chromosomes, but that there were other lower chromosome numbers present. Root-tip squashes for the eight plants from primocane A showed that all cells had 31 chromosomes, and root-tips of the seven plants from primocane C had cells with 29 chromosomes only. Thus, propagation of plants with the high and unstable chromosome number of primocane B resulted in plants which were still mitotically unstable, whereas propagation from primocanes A and C of lower chromosome number, resulted in plants which were mitotically stable.

A few isolated cells have been seen in root-tips of some unstable plants with as few as 10 or 11 chromosomes—below the gametic number of either parent. However, when propagations of root-cuttings or root suckers have been made of the plants in which these low chromosome numbers were observed, the plant with the lowest chromosome number obtained had 26 chromosomes.

Cytological Observations and Discussion

Cytological studies of root-tip squashes from the 6x seedlings of selfed Eldorado showed many cells with fewer than 42 chromosomes (Table II). In addition there were seen configurations which suggested how these lower chromosome numbers could have been derived from pre-existing higher numbers. The cytological observations on all of the other mitotically unstable plants have been closely comparable to those on the 6x seedlings of Eldorado. Accordingly, it is considered that the same sequence of mitotic events occurs in all of these plants. Admittedly any sequence of mitotic events is hypothetical when it is based on observations of fixed material, since there is no way of knowing that each stage or phase gives rise to the next. However, the sequence presented gives a logical order for the configurations seen and does allow for the observed diminution of chromosome numbers.

The critical observations on which this sequence is based are:

  1. At metaphase there may be two (or more) distinct plates of chromosomes each with its individual spindle (Figure 8F, G). The total number of chromosomes in the two plates is equal to the number of chromosomes in some of the neighboring cells. Although it has been possible to make these chromosome number determinations with such precision in relatively few cases, it has always been strikingly apparent on the basis of the bulk of the chromosomes (cf. prophase and three groups of chromosomes in prometaphase in Figure 8B).
  2. At anaphase there are four (or more) groups of chromatids with two spindles (Figures 8H, 9A and B).
  3. At telophase there are four (or more) groups of chromatids with phragmoplasts evident at right angles to the orientation of the spindles and equidistant between the chromatid groups. Phragmoplasts have only been seen in the region where there was a spindle.

The stages of the cytological mechanism resulting in the decrease of chromosome numbers are considered to occur in the following sequence:

  1. At prophase there is either an apparently single prophase nucleus or there is evidence of an irregular grouping of the chromosomes. In the example shown in Figure 8A there was evidence of two distinct groups. This stage is not considered critical because of the possibility that such apparent groupings are caused by the squash technique. However, the groups of chromosomes with independent spindles at metaphase and anaphase give no evidence of being caused by squashing.
  2. The chromosomes at prometaphase are often more clearly seen in definite groups. For example, in Figure 8B three such groups are shown and in Figure 8C the more common occurrence of two distinct groups is illustrated. These groups were either of equal numbers of chromosomes or were quite unequal.
  3. At metaphase there are usually two distinct plates of chromosomes as shown in Figure 8D. These plates are oriented in various ways. Some of the orientations are shown in Figures 8D (polar), 8E (one polar, one side view), 8F (two plates in side view and inclined), and 8G (two plates in side view and perpendicular). In order to conform to the terminology of Wilson et al.16 and Vaarama13 this stage is termed a split metaphase. It is used in the sense that the chromosomes of a single cell are separated into two distinct or split groups, each with its own spindle. There seems to be no reason to consider the spindle to be split into two parts. The term "double-plate metaphase"12 might have been used equally well.
  4. A split anaphase follows the split metaphase. At this stage there are usually four groups of chromatids with two spindles (Figures 8H, 9A and B). In the more uncommon event of having three groups of chromosomes at prometaphase, there are then six groups of chromatids and three spindles at anaphase. The number of chromatids going to each anaphase pole is dependent on the number of chromosomes that were in each group of the split metaphase. The orientation of the spindles is dependent on the orientation of the plates at split metaphase.
  5. At telophase phragmoplasts are evident between the groups of chromatids in the region of the spindles and no phragmoplast is formed between the original or split groups of chromosomes where there is no spindle (Figure 9C). After this stage there are formed three cells, two of which have a smaller chromosome number than at prophase and one cell which is binucleate. It should be noted that a telophase from a split metaphase such as shown in Figure 8F could have two groups of chromosomes in close proximity. If these groups were enclosed by a single nuclear membrane, then a partially binucleate or lobed nucleus would result. Binucleate cells (Figure 9D) and cells with lobed nuclei were of frequent occurrence in all the mitotically unstable plants.
TABLE I. Parentage and numbers of mitotically unstable seedlings   TABLE II. Chromosome numbers from 32
root-tips of (6x) seedlings of Eldorado selfed*
Parentage Numbers of
of cells
42 16
41 4
56-49 Eldorado (4x) X self 5 37 1
56-53 Austin Thornless (8x) X (8x) Eldorado 1 35 3
56-56 Johnson (ca. 8x) X (8x) Eldorado 2 34 3
56-58 Wild seedling (ca. 6x) X (8x) Eldorado 1 32 4
56-61 Young (7x) X (8x) Eldorado 5 30 2
56-62 (8x) (Merton Thornless X Brainerd) X self 1 29 3
56-63 (8x) (Merton Thornless X Brainerd) X (8x) Eldorado 1 28 12
5676 Cascade (9x) X (8x) Eldorado 2 27 2
56-79 Boysen (7x) X (8x) Eldorado 3 26 1
56-82 Hailsham (4x) Raspberry X Austin Thornless (8x) 5 25 1
  Total 28 24 3
      *Only those chromosome number
determinations within an error of one.


Figure 7
Three primocanes which originated as root suckers of a single unstable hybrid (Young X (8x) Eldorado). Primocanes are lettered from left to right, A, B and C See text for details.

Figure 9G and H illustrates two cells from one root-tip with very different chromosome numbers. In one cell there are 11 chromosomes and in the other 43. The root-tip was taken from, a plant of the cross Cascade (9x) (2n = 63) X (8x) Eldorado (2n = 56). Chromosome number determinations made from axillary leaf-bud squashes, as well as root-tip squashes early in the life of the seedling showed most cells to have 59 ± 2 chromosomes. Each of the mitotic stages of split divisions was seen in this plant. Accordingly, the sequence of events leading to the decrease in chromosome numbers from 59 to II is considered to have occurred by the cytological mechanism outlined above.

Most of the unusual configurations seen in mitotically unstable plants could satisfactorily be classified in the afore mentioned sequence of mitotic events. However, a few metaphases were seen which were "T"-shaped (Figure 9E), or "X"-shaped (Figure 9F) or even "V"-shaped. With configurations such as these, only one spindle (Figure 9E) in each cell has been seen. These observations differ from those of Vaarama13 who classified these metaphases with split metaphases. The inability to effectively demonstrate the presence of two spindles in each cell in "T", "X", and "V"-shaped metaphases suggests that the separation of chromosomes into split groups is not complete in these configurations. These abnormal metaphases would be followed by anaphases with larger numbers of chromosomes at one pole than the other. A few such anaphases have been observed. This phenomenon would be an additional explanation for the variability in the chromosome numbers observed.

Most other features of the diminution of chromosome numbers presented in this paper are very similar to those described by Vaarama13. In both studies comparable mitotic stages including separate or split spindles were found, whereas no multipolar spindles were seen. However, Vaarama13 concluded that it was the colchicine which was responsible for the upset in the spindle mechanism even in the second generation (C2). That colchicine is not necessary to cause mitotic instability is evident, since the (6x) Eldorado seedlings reported here were from a progeny of untreated selfed Eldorado. A further difference between these studies is that many binucleate cells were seen in the present work. The presence of many binucleate cells obviates for Rubus. Vaarama's postulate for Ribes that "Evidently cell wall formation is in principle similar to tetrad formation after the second meiotic division." Cell walls were formed only at right angles to the spindles, and were not formed between split plates where there was no spindle.

Another point of variance between these studies concerns the interpretation of the presence of lobed or constricted nuclei. Vaarama suggests that these are evidence for a certain degree of autonomy of nuclear parts and that they arise from single resting nuclei. However, two split metaphase plates which are inclined towards each other (Figure 8F) could give rise to a lobed nucleus at telophase (Huskins and Cheng, Figure 6I7). Accordingly, lobed nuclei in this study are considered to result from, split divisions rather than to represent forerunners of them.

Snoad11 found cells with two distinctly separate spindles in Hymenocallis, as well as tripolar spindles. He terms these phenomena "spindle abnormalities." This would seem to be an error in emphasis when considering two separately organized spindles which are perfectly normal in appearance and function, as seems to be the case in Rubus.

Menzel and Brown9 discuss mosaic formation and somatic reduction in Gossypium and conclude that one possible mechanism for a common cause of these phenomena would be "some type of atypical segregational mitosis." Mosaic plants were found by them between the 4x (2n = 52) and 8x (2n = 104) levels of ploidy, but not at the 2x—3x levels (Table 2, loc. cit.) In Rubus, plants exhibiting mosaicism were found also, and all these plants were mitotically unstable. This instability was found to occur above the 2n = 42 (6x) level. Menzel8 reports another case of somatic reduction in a mosaic-forming hybrid of Gossypium. Here the somatic chromosome number had decreased from 78 to 69. It is probable that the same general phenomenon is involved in both Gossypium and Rubus.

Wilson et al.16 have described variations of mitosis in both treated and untreated onion root-tips. The unusual divisions found by them in untreated root-tips closely approximate those reported here in Rubus. The mechanism causing the spontaneous reduction of chromosome number reported by Huskins and Cheng7 could be interpreted as being of a similar nature also.

It should be noted that Huskins6, Huskins and Cheng7, Wilson14, Wilson and Cheng15, and Wilson et al.16 were all working with diploid plants in which a very low percentage of cells with decreased chromosome numbers were seen. These cells with sub-diploid numbers would be unlikely to divide again. In both polyploid Gossypium and Rubus many of the cells resulting from irregular mitoses are capable of further division. For example, in Gossypium, Menzel8 found an entire mutant branch, and in the present study of Rubus over 100 plants have been propagated asexually that had chromosome numbers less than those of their respective mother plants. In Rubus, extremely low numbers of 10 and 11 chromosomes have been seen rarely, but as was pointed out, it has not been possible to obtain asexually propagated plants with such chromosome numbers.

The mechanism causing mitotic instability presented here is considered too inexact to use the terms "reductional" or "segregational". It is inexact in that all gradations, from two completely separate metaphase plates with separate spindles, to "T"-shaped metaphases with one spindle, to a single metaphase plate with one or two chromosomes off the plate, have been seen. However, when there are two separate spindles in a single cell, the end result is the production of cells with a lower chromosome number. It is only in this sense that the division is reductional. The fact that two groups of chromosomes separate from a single resting nucleus permits the possibility of the expression of some factors that were previously masked by dominance (i.e., segregation). However, when considering high  (7x-8x) polyploids, the most likely result would be the loss of factors still represented in the chromosomes retained. Thus, the majority of the genetic changes from the split divisions would be quantitative rather than a qualitative or segregational nature.

The lack of phenotypic differences in the plants propagated by leaf-bud cuttings in contrast to the great phenotypic differences between the plants propagated by root-cuttings may be explained by contrasting the complex structure of a vegetative shoot apex with the simple structure of a root apex. The vegetative shoot apex in Rubus consists of three histogenic layers. Each of these three layers is maintained by the division of 3-5 primary cells at the extreme apex3. Unless one or more of these primary cells undergoes a split division, the growing point will continue to produce cells with the same chromosome number. The axillary leaf-bud primordia are in close proximity to the apical dome and originate from very few cells. Consequently there are few cells in this restricted area in which it is possible for split divisions to occur, in contrast to the multitude of dividing cells below the apex of the dome. Split divisions occurring later in the ontogeny of leaves have caused the formation of mosaic patterns and in several cases, of a whole leaflet of different morphology. On the other hand, the roots are a many branched system having only one histogenic layer. The secondary roots in this system arise adventitiously from the primary root. Split divisions occurring in the primary root will give rise to isolated groups of cells with lower chromosome numbers than the zygotic number. If then, secondary roots arise from isolated groups, the lower chromosome numbers would be perpetuated. Further split divisions and branching of roots would result in further decreases of chromosome numbers.

Thus from mitotically unstable plants it has been possible to isolate and propagate aberrant forms by means of root-cuttings that would not be apparent from propagation by leaf-bud cuttings. It follows that leaf-bud propagation is a more reliable means of propagating unstable clonal material than are root-cuttings.

Mitotic instability is a phenomenon that causes difficulties for the plant breeder who may wish to maintain a naturally occurring polyploid such as a (6x) seedling of Eldorado, or an artificially induced polyploid such as (8x) Eldorado. Since the canes are biennial in Rubus, it is necessary to propagate such unstable plants as (8x) Eldorado by leaf-bud cuttings each year in order to have any hope of maintaining the desired level of ploidy for crossing in the following year. The difficulty of maintaining certain high polyploids at a known level of ploidy has proved to be a real one in the program at Maryland. It has been necessary to continually propagate and reexamine cytologically many of the polyploids that are used in the breeding program.

It is extremely difficult to state with any assurance that one Rubus plant is completely mitotically stable and another is not. If a large number of split divisions are found it is assumed that the plant is mitotically unstable. However, if no split divisions are found and the chromosome number appears to be constant, there is always the possibility that insufficient observations have been made. Nevertheless, certain of the progeny from the crosses presented in Table I appeared to be quite stable, although considerably above the 6x level of ploidy. The genetic constitution necessary for mitotic instability, or conversely for mitotic stability, is not evident. However, one factor involved may be the degree of autopolyploidy of the unstable plant. This is indicated by the facts that (4x) Eldorado is mitotically stable, (6x) seedlings of selfed Eldorado are unstable, and (8x) Eldorado is markedly unstable; whereas apparently an allopolyploid variety4—such as Cascade (9x) is stable. In this connection one might speculate whether this is the reason that no eastern erect blackberries occur naturally above 4x in ploidy, whereas the western trailing blackberries are found to the 12x level of ploidy.

Many of the crosses in Table I had one parent which was an artificial polyploid, such as (8x) Eldorado. It should be noted that the nature of the breeding program was such that (8x) Eldorado and other artificial polyploids were involved in many crosses. Accordingly, it is not possible to say that artificial polyploids are the ones most likely to be involved in the production of unstable plants. Nevertheless, it seems that the plant breeder should be particularly watchful for mitotic instability when working with synthesized polyploids.8,9,10


Twenty-eight mitotically unstable seedlings were found among 10 progenies of Rubus.

The number of chromosomes in some cells was found to range down to a number less than the gametic number of either parent. The essential feature of the mechanism of this diminution was the grouping of the chromosomes of one cell into two plates, each with its own spindle at metaphase. Studies by others indicate that a decrease of chromosome number is not peculiar to this genus. Similar, or perhaps identical phenomena have been described for Ribes, Hymenocallis, the Triticinae and Gossypium. However, mitotic instability in Rubus is more than a cytological anomaly, as is shown by the results of asexual propagation. Root-cuttings have given rise to plants with differing phenotypes. Cytological studies of these plants showed that they had low chromosome numbers which were usually aneuploid and that they were mitotically stable. On the other hand, leaf-bud cuttings produced plants with a similar phenotype to the mother plant and perpetuated the mitotic instability. The differences between the results from root-cutting propagation and leaf-bud cutting propagation are explained by the contrasting histogenesis of roots and shoots.

Although colchiploid parents are not necessary to cause mitotic instability in the seedlings, it appears that they are often involved. The decrease in chromosome number of mitotically unstable Rubus stocks is a plant breeding problem which may be overcome by annual asexual propagation of the top of the plant by leaf-bud cuttings.

Literature Cited

  1. BRITTON, DONALD M. and J. W. HULL. Mitotic instability in blackberry seedlings from progenies of Boysen and of Young. Jour. Hered. 47:205-210. 1956.
  2. EINSET, JOHN. Apomixis in American polyploid blackberries. Amer. Jour. Bot. 38: 768-772. 1951.
  3. ENGARD, C. J. Organogenesis in Rubus. Univ. Hawaii Res. Pub. No. 21. 234 pp. 1944.
  4. FISCHER, H. E., GEORGE M. DARROW and G. F. WALDO. Further chromosome studies of some varieties of blackberries. Proc. Amer. Soc. Hort. Sci. 38:401-404. 1941.
  5. HULL, J. W. and DONALD M. BRITTON. Early detection of induced internal polyploidy in Rubus. Proc. Amer. Soc. Hort. Sci. 68: (in press). 1957.
  6. HUSKINS, C. LEONARD. Segregation and reduction in somatic tissues. I. Initial observations on Allium cepa. Jour. Hered. 39:311-325. 1948.
  7. —————————— and K. C. CHENG. Segregation and reduction in somatic tissues. IV. Reductional groupings induced in Allium cepa by low temperature. Jour. Hered. 41:13-18. 1950.
  8. MENZEL, MARGARET Y. Polygenomic hybrids in Gossypium. III. Somatic reduction in a phenotypically-altered branch of a three-species hexaploid. Amer. Jour. Bot. 39: 623-633. 1952.
  9. —————————— and META S. BROWN. Polygenomic hybrids in Gossypium. II, Mosaic formation and somatic reduction. Amer. Jour. Bot. 39:59-69. 1952.
  10. SACHS, LEO. Chromosome mosaics in experimental amphiploids in the Triticinae. Heredity 6:157-170. 1952.
  11. SNOAD, BRIAN. Somatic instability of chromosome number in Hymenocallis calathinum.  Heredity 9:129-134. 1955.
  12. UPCOTT, A. The nature of tetraploidy in Primula kewensis. Jour. Genet. 39:79-100. 1939.
  13. VAARAMA, A. Spindle abnormalities and variation in chromosome number in Ribes nigrum. Hereditas 35:136-162. 1949.
  14. WILSON, G. B. Cytological effects of some antibiotics. Jour. Hered. 41:227-231. 1950.
  15. —————— and K. C. CHENG. Segregation and reduction in somatic tissues. II. The separation of homologous chromosomes in Trillium species. Jour. Hered. 40:3-6. 1949.
  16. ——————, MARY E. HAWTHORNE and TE MAY TSOU. Spontaneous and induced variations in mitosis. Jour. Hered. 42:183-189. 1951.