Hereditas 25: 33-47 (1939)
DIFFERENTIAL POLYPLOIDY WITHIN THE BLACKBERRIES

BY AKE GUSTAFSSON
INSTITUTE OF GENETICS, SVALOF

SPECIES DIFFERENTIATION.

ALL European Rubi Eubati can be divided into two large groups: Eubati veri and Rubi Corylifolii. Eubati veri (as well as Rubi Corylifolii) consist of a series of species and microspecies with different taxonomic value. From a practical point of view they may be placed in six taxonomical groups. The first group comprises the very few main-species, which are sexual and diploid and have a southern area of distribution, namely R. tomentosus and ulmifolius. Both are plainly distinguished from other Rubi Eubati. The first-mentioned species forms a uniform and seemingly depauperated group of biotypes, while R. ulmifolius consists of a large number of forms closely related to each other. In spite of this polymorphy R. ulmifolius is a characteristic and isolated species. In the same group of species FOCKE [(1878 and) 1914] also placed R. caucasicus and incanescens, which so far have not been examined cytologically or genetically. Even R. caesius — distributed throughout Europe and mostly rather common — is placed by him in this group. LIDFORSS (1914 and earlier) showed, however, that Scandinavian members of this species are obligate apomicts, and GUSTAFSSON (1933a and unpublished; cf. also LONGLEY, 1924 and ROSANOVA, 1934) determined that several representatives have the tetraploid chromosome number.

Within the remaining Eubati veri there is a great variation due to hybridization. In fact, the different groups (subsectios) consist of a large number of transitional forms. Great many localized and endemic microspecies have arisen. Except a few species with a large area of distribution, the different regions display clusters of forms — often vegetatively propagated in a luxuriant manner — which it has been impossible to group together with the non-endemic species occurring in the region. But as has been shown by LIDFORSS, even several so-called good species are heterozygous to a surprisingly great extent. Crosses between them or even segregation products may therefore give origin to floras of blackberries, so rich in biotypes that each limited area con­tains its own endemisms. This is especially true of the so-called Rubi Glandulosi, consisting of hundreds of microspecies grouped around R. hirtus and R. serpens (the pentaploid R. Bellardii also belongs to this group; GUSTAFSSON, 1933b). It is as difficult to classify these microspecies as those of the Rubi Corylifolii. While these have their maximum of variation in northern Germany and southern Scandinavia, Rubi Glandulosi have a more southern type of distribution, with their greatest polymorphy in southern France, in Switzerland and Hungary. An astonishing fact is their pronounced capacity of migration into high-alpine areas (GUSTAFSSON, 1933b).

Rubi Corylifolii are — as has been conclusively proved by LIDFORSS — primary and secondary hybrid and segregation products of crosses between Eubati veri and the tetraploid R. caesius. In most cases it is impossible to elucidate with certainty which Eubati veri have taken part in these processes of hybridization. Probably in many cases, the Corylifolii occur in areas where the corresponding Eubati veri do not exist. Due to this panmixis a satisfactory phylogenetic and taxonomic treatment of the Corylifolii is impossible.

Rubi Eubati can be divided into the following six classes of varying taxonomical value.

  1. Isolated, sexual species (cf. above).
    II — VI: ± apomictic biotypes and biotype compounds.
  2. Microspecies, which are taxonomically uniform and have a total-European area of distribution.
  3. Microspecies, which form centres in larger complexes and have a large area of distribution.
  4. Microspecies, which are isolated and taxonomically distinguishable but have small areas of distribution.
  5. Endemic microspecies, taxonomically isolated.
  6. Microspecies, which are localized and cannot be characterized taxonomically, mostly with transitional forms to other microspecies and compounds.

Eubati veri form part of the species groups I-VI, Rubi Corylifolii of the groups IV—VI.

There is no correlation between the taxonomical value of the microspecies and their chromosome numbers. Owing to the fact that polyploids with an odd number of genomes (3x, 5x) do not hybridize easily or in such a case give cytologically heterogeneous and non-viable offspring, a few of these polyploids tend to become isolated.

This grouping of Rubi Eubati into different classes according to their distribution, uniformity (phenotypical constancy) and systematical distinctness is rather arbitrary. Therefore a parallel to the different units of the sexual species (TURESSON, 1929 and earlier) cannot be drawn. The two sexual species, R. tomentosus and R. ulmifolius, are presumably two coenospecies. Most of the apomictic microspecies of groups II-VI (V) can, however, neither be equalled to coeno- or ecospecies, nor be regarded as ecotypes. Of special interest is the fact that the two microspecies, R. thyrsanthus and candicans (belonging to group IV) with the triploid chromosome number 3x=21, can be compared with climate ecotypes in the sexual species. Their areas of distribution do not agree. R. thyrsanthus has a more eastern, R. candicans, on the contrary, a more western distribution. Taxonomists have often regarded these two microspecies as hybrids of R. tomentosus. Their chromosome numbers favour this view (n=7 x n=14 -> 2n=21). The transitional forms in the boundary zone of the two areas are not necessarily crosses between the two microspecies (note the triploid chromosome number), but may be original hybridization products equal to the microspecies themselves.

As has been shown above, it is possible to search for single biotypes or biotype compounds in the great variety of hybrid and segregation products which keep isolated and for that reason can be treated systematically. However, all transitions in systematical value exist between these few microspecies, which can be classified without difficulty, and those which are groups of microspecies or groups of biotypes transitional to each other rather than different species. This fact renders the taxonomical treatment of the genus arbitrary and superficial. This has also been shown by the occurrence of the many types of monographs published. FOCKE, who studied the genus for about 50 years, finally adhered to a summary treatment. Microspecies, resembling each other, were placed together in groups, often in an artificial manner. Among the hundreds of microspecies, already described, most of those which were extremely localized, poorly known or badly described were left out, as well as most of the transitional forms connecting two or more microspecies.

So far as is known, apomixis occurs only in the subgenus Eubatus, but not in all groups. The South American representatives have not hitherto been examined experimentally, only one of them (R. bogotensis) cytologically. This species has the diploid chromosome number (2n=14). The North American species, several of which are closely related morphologically to the European Suberecti, are always sexual with the exception of course of introduced species or hybrids of these (BRAINERD and PEITERSEN, 1921; GUSTAFSSON, 1930; DARROW, 1937). LONGLEY (1924) has examined many species and found polyploidy to be represented abundantly. In spite of their sexuality the North American species often form biotype compounds with transitions to each other which are difficult to classify.

THE DIFFERENT POLYPLOIDY WITHIN THE EUROPEAN BLACKBERRIES.

The facts mentioned above show that polyploidy is a common phenomenon among the Eubati veri. All apomictic microspecies are triploid, tetraploid or pentaploid. Of 39 examined microspecies from 16 different localities the number of representatives in the different groups of polyploidy is 3, 32 and 4 respectively. The tetraploids are plainly predominant (82%). Hexaploids or microspecies with a still higher chromosome number do not occur, not even in the form spheres which are most heterogeneous and thus probably also most heterozygous. Nor do the endemic microspecits show a higher degree of polyploidy. Five out of 33 Scandinavian microspecies are endemic (R. subvelutinus, R. confinis, R. taeniarum, R. vestervicensis, R. kollundicola). The chromosome numbers are 21, 28, 28, 35, 35 respectively. Neither are the most northern representatives particularly high-polyploid. The most northern microspecies of the Swedish east coast and inland are R. suberectus, R. sulcatus, R. plicatus, and R. thyrsanthus, the first three being tetraploid, the last triploid. The most common microspecies along the west coast of Norway are R. suberectus, R. plicatus, R. Selmeri, and R. fissus, all tetraploid.

The few pentaploid microspecies are of particular interest. Two of them, R. kollundicola and vestervicensis, are localized microspecies. R. cordifolius was described from a small region. Later authors have widened its area of distribution but at the same time also changed the description. All are vegetatively luxuriant and apparently vital. Their small areas of distribution may therefore be attributed either to a late origin or to a lack of apomictic propagation (the first explanation, though not established, seems to be the more plausible one). R. Bellardii, on the other hand, is a powerful microspecies with a total-European area of distribution and a high constancy (cf. GUSTAFSSON, 1933b).

Rubi Corylifolii have the chromosome numbers 28, 35, 42, ±45, 49. Since 1933, 85 representatives have been examined and these correspond to ±74 different types. In the different classes of polyploidy 2n=28, 35, 42 and 45-49, respectively 42, 13, 17 and 2 types occur. Polyploids with a higher number than 28 are about as common as tetraploids. Corylifolii have therefore a greater tendency than Eubati veri to form high-polyploids. Only the chromosome number 28 implies an equilibrium in the Eubati veri, while among the Corylifolii the number 28 as well as 35 and 42 purport equilibria (57, 18 and 26% respectively). The hexaploid Corylifolii occur exclusively in the western parts of the investigated region (cf. below). The western Corylifolii show the following numbers of types in the different classes: 2n=28 (27 repr.), 2n=35 (7 repr.), 2n=42 (17 repr.), 2n=45-49 (2 repr.). The percentages are 51, 13, 32 and 4. The surplus of tetraploid types has thus been still more diminished.

This difference in degree of polyploidy between the European Eubati veri and the Corylifolii cannot be the result of an unequally intense hybridization in the two groups. As mentioned above, Eubati veri form populations of hybrids and segregations, which in certain regions are as difficult to classify as the Corylifolii. Certainly, as yet only a few representatives belonging to Rubi Glandulosi have been examined cytologically (5 repr.) but all of them — except R. Bellardii — are tetraploid. According to MUNTZING (1936) the varying viability in different classes of polyploidy is due to a chromosome optimum, above which the polyploid zygotes become non-viable or sublethal. The pentaploid Eubati veri are however — in spite of their rareness — vegetatively luxuriant and fertile. One of them, R. Bellardii, possesses a large area of distribution and probably a very great rate of spreading. Several plant genera are known in which polyploidy does not arise at all. There must be processes other than that mentioned above which prevent the origin of high-polyploidy or in certain plant genera even the occurrence of triploidy or tetraploidy. As the capacity to form restitution nuclei and unreduced gametes in certain parthenogenetic groups is genotypically determined (GUSTAFSSON, 1935a), so the (pre)meiotic formation or the postmeiotic function of unreduced gametes may depend upon the genotypical constitution. R. Bellardii, which has a disturbed meiosis, never forms dyads or triads. Pentads, hexads, heptads etc. occur, however, rather frequently. (Cf. also BLAKESLEE and BUCHHOLTZ, 1929 and SCHLÖSSER, 1936).

An interesting fact is that the American Eubati veri are not only total-sexual but that in spite of their sexuality they have a much stronger tendency than the European apomicts to form high-polyploids. LONGLEY (1924) investigated 25 American and 3 European Eubati. Of the latter, R. thyrsoideus (a group name for R. thyrsanthus and related forms) had 2n=21, R. corylifolius and R. caesius var. turkestanicus 2n=28. Of the 25 American Eubati veri 6 were diploid, 12 triploid, 2 pentaploid, 4 hexaploid, 1 octaploid. Apomixis and a high degree of polyploidy are therefore not necessarily simultaneous phenomena. The European Eubati veri must contain a physiological or genetical factor which prevents the increase of polyploidy or — from the other point of view — the American Eubati veri and in certain crosses R. caesius must contain a character increasing polyploidy.

Very interesting in this connection is the spontaneous hybrid R. idaeus x caesius. It is highly sterile and lacks the capacity of apomictic propagation (pseudogamy a recessive phenomenon). Individuals, which appear morphologically to be back-crosses to caesius and idaeus, are not rare. From a theoretical point of view the primary hybrid should possess the triploid chromosome number (n=7 x n=14 -> 2n=21). This chromosome number explains the sterility in the absence of apomixis. Of 5 primary hybrids and back-crosses two possessed the triploid chromosome number (Rubi Scand. Chrom. Exam. Nos. 148 and 172), one the number 28 (147), one the number 35 (270) and one the number 42 (127). This hybrid has thus the capacity of forming polyploid biotypes in a high degree. ROSANOVA (1934) examined the artificial F1-hybrids between R. idaeus and caesius and found in one offspring the number 28 (presumably n=14 x n=14) and in one offspring the number 42 (presumably n=14 x n=28). Most probably these data indicate the function of unreduced R. caesius-genes. Why these do not function within the species itself or its crosses with certain Eubati veri and Corylifolii, cannot be answered. However, the lack of increase of polyploidy in the European Eubati veri does not depend upon a higher degree of fertility in the crosses between them for sterile hybrids (respectively back-crosses or segregations) are not infrequently produced in nature.

Similarly, among the spontaneous Corylifolii types occur which show a high degree of sterility. Each limited blackberry-flora contains a number of such products. One of the most interesting Scandinavian representatives in this respect is R. glaucovirens, which stands morphologically between R. Wahlbergii and R. caesius. Its area of distribution is comparatively large: Blekinge, Bornholm, Skåne, Sjålland. In the east of Skåne and in Blekinge it is common. The chromosome number is 4x=28. Despite this balanced number it is highly sterile. The well-developed flower-stands carry only a few fruits and at most one or two ripe carpels per flower. Since R. glaucovirens is very uniform, its large area of distribution cannot be explained by the occurrence of parallel crossings in different parts of the area. Similar types with a higher chromosome number have not been found. All pentaploid biotypes of the area in question belong to quite different microspecies (mostly to R. pruinosus) and hexaploid biotypes have not arisen, either in the east of Skåne, Blekinge or Bornholm. The few seeds have probably without exception arisen apomictically, and their small number is sufficient to explain the great distribution of the microspecies.

THE DIFFERENTIAL POLYPLOIDY WITHIN THE CORYLIFOLII.

Although the number of examined Rubi Corylifolii amounts to only 100 representatives, and many gaps in our knowledge must be filled, definite conclusions can be drawn in certain respects.

Of 34 examined individuals from the eastern Rubus-area, representing ±25 different types, 33 possessed the tetraploid or pentaploid chromosome number, none was hexaploid and one heptaploid. The equilibrium is undoubtedly 2n=4x. Outside the R. pruinosus-complex only three representatives were pentaploid. Controls of the pentaploid number of R. fioniae from Ostergotland must be secured from new material. The two remaining types are localized (the pentaploid from Bornhoim inadequately known). An interesting fact is that R. pruinosus always is pentaploid in the eastern as well as in the western Rubus-area in spite of its pronounced polymorphy. Taxonomists have often regarded R. pruinosus as a hybrid between R. idaeus and Rubi Corylifolii. The pentaploid chromosome number can theoretically arise, if unreduced Corylifolii-gametes (4x=28) are fertilized by or fertilize reduced R. idaeus-gametes (x=7). With this explanation several difficulties disappear. Apparently it is impossible to explain the polymorphy of R. pruinosus by the origin of mutations (embryo-mutations). Nor can segregations of R. pruinosus explain the great number of different types. Different chromosome numbers should then have arisen. Also the fact that (back) crosses between R. idaeus and R. caesius often resemble R. pruinosus, argues in favour of the opinion of taxonomists. Plant No. 127 in Rubi Scand. Chrom. Exam, has a strong resemblance to R. pruinosus. The hexaploid chromosome number distinguishes it from all R. pruinosus-types. If this explanation is correct, unreduced gametes must now and then arise in certain crosses also among the east Swedish Corylifolii. Finally it may be mentioned that LIDFORSS in his experiments obtained types resembling R. pruinosus from the cross R. thyrsanthus x R. caesius.

During the last years it has been shown that different chromosome races of a species (or species closely related to each other but with different chromosome numbers: ecospecies; TURESSON, 1929) have often very dissimilar distribution, ecology, life form (MUNTZING, 1936; TISCHLER, 1935). The changed properties, brought about by the increased chromosome numbers (often through autopolyploidy) show themselves to be mainly quantitative — in size, winter-hardiness, amount of vitamins, and so on. These changed properties permit the coenospecies to widen its area of distribution. The environment does not change the chromosome numbers (HAGERUP, 1931) but acts only passively through selection among biotypes with different chromosome numbers.

Even purely historical factors and factors due to immigration phenomena must be of very great importance. This is completely proved by the difference with regard to degree of polyploidy in the subatlantic and the middle Baltic areas. (The differential polyploidy may even serve as a factor separating these two plant-geographical regions; GUSTAFSSON, 1935b). For nobody can maintain that the subatlantic or atlantic climates are able to produce high-polyploids by themselves, while, the middle Baltic climate lacks this ability. Nor is there any reason to assume that high-polyploids grow better in the western than in the eastern area. The cause of the differential polyploidy must lie, therefore, in the fact that the immigrants to the eastern area consisted and still consist of biotypes and biotype compounds lacking the power of increasing the chromosome number above the pentaploid status. The immigrants to the western area, on the other hand, possessed this capacity.

Several microspecies show different chromosome numbers in the eastern and in the western areas. This is especially true of R. Wahlbergii and R. ambifarius (coll.). Seven representatives of the first microspecies have been examined (in the investigations before 1933 also two other individuals). The chromosome numbers are the following: East-Sweden 28, ±30, (28), West-Sweden + Norway 28-29, 35, 35, 35 (42), Denmark 35. The related species, R. vexatus, has 6x=42. R. ambifarius gives for the two areas respectively 28, (28), and 28, 35, 42.

R. Mortensenii, growing in the western region, shows the numbers 28, 27-28, 28, 42, R. eluxatus, also in the western region, 28, and in an earlier investigation ±45. An individual phenotypically related to R. eluxatus (Rubi Scand. Chrom. Exam. No. 175) revealed also the number ±45. R. leiocarpus gave the numbers 28, 49. R. nitens behaves in the same way, possibly also R. ciliatus and R. tiliaceus. R. nitens is a polymorphous microspecies. The existence of two different chromosome races (28, 42) is therefore not difficult to interpret. Against these numerous cases of chromosome races in the western region only one sure case from the eastern region can be mentioned: two R. gothicus-types from Bornholm have 2n=28 and 35. Chromosome races containing the hexaploid numbers have never arisen in the eastern region.

These cases of differential polyploidy within a microspecies prove beyond doubt the view that in the western area phenomena exist which produce high-polyploids and change low-polyploids, immigrated from other regions, into high-polyploids. Similar Corylifolii may of course have arisen from different crosses, so that the eastern tetraploids are not necessarily the ancestors of the resembling western penta- and hexaploids but only parallel products. The western Corylifolii cannot possibly have arisen from the eastern biotypes exclusively through autopolyploidy, because autopolyploids would then have arisen in the eastern region itself. Since the representatives of a microspecies, containing different chromosome races, are not merely superficial in resemblance but are almost identical phenotypically (R. Walhbergii, partly R. ambifarius), affinity in a phylogenetical sense is probable. This, is obvious when high- and low-chromosomal types occur in the same region (R. leiocarpus, R. Mortensenii, R. nitens, partly R. Wahlbergii).

SUMMARY.

  1. European apomictic Eubati veri are triploid, tetraploid or pentaploid but never hexaploid or heptaploid. 82% of the apomictic microspecies are tetraploid.
  2. American Eubati veri (sexual) are diploid, triploid, pentaploid, hexaploid and octaploid.
  3. Rubus idaeus x caesius easily forms unreduced gametes. Therefore the F1 and back-crosses exhibit the triploid, tetraploid, pentaploid and hexaploid chromosome numbers. R. caesius is capable of forming unreduced gametes as father-parent in the cross R. idaeus x caesius. Nevertheless, so far as is known, R. caesius itself is always tetraploid in nature.
  4. East-Scandinavian Rubi Corylifolii are mostly tetraploid. No hexaploids occur. Most pentaploids belong to R. pruinosus coll.
  5. West-Scandinavian Rubi Corylifolii are tetraploid, pentaploid, hexaploid, superhexaploid, heptaploid.
  6. Several eastern Corylifolii increase their chromosome numbers when migrated to the western region. Several western Corylifolii contain different chromosome races, often with the tetraploid or pentaploid + the hexaploid or heptaploid number.
  7. The differential polyploidy cannot be due to a different degree of heterozygoty or sterility, nor to a selectional effect of climate. Neither is there any possibility of the subatlantic climate directly producing polyploids with high chromosome numbers.
  8. The differential polyploidy within the Corylifolii is probably due to an increase of the chromosome number in crosses between R. caesius and atlantic Eubati veri or to crosses between atlantic Corylifolii (which themselves have arisen from crosses between R. caesius and Eubati veri) and to status quo in crosses between R. caesius and eastern Eubati veri or between eastern Corylifolii.

The author is indebted to Dr. B. P. KAUFMANN, Carnegie Institution of Washington, Cold Spring Harbor, U. S. A., for some corrections of the manuscript.

EUROPEAN EUBATI VERI.

Rubi Scandinavici, chromosomi examinati.

The author began some years ago to collect material for a parallel cytological and taxonomical investigation of the genus Rubus. This exsiccate is made in cooperation with Mr. C. E. GUSTAFSSON, Trelleborg, Sweden, and when finished will be deposited in the Botanical Museums of Lund and Stockholm.

Triploid species.
2n=21.

Tetraploid species.
2n=4x=28.

Pentaploid species.
2n= 5x=35.

Chromosome numbers of other Eubati veri, examined by the author.

Diploid species.

Triploid species.

Tetraploid species.

RUBI CORYLIFOLII.
I. The East of Sweden.
1. Blekinge-Uppland.

    No.  
{ R. Wahlbergii ARRH. 151 2n=4x= ±30
R. Wahlbergii ARRH. v. tenuifolius ARESCH. 4 28
R. nemorosus ARRH. v. suberiocarpus ad int. 86 28
R. nemorosus ARRH. v. subglandulosus LUND 115 28
  R. gothicus FRIDER. p. p. 84 28
{ R. centiformis FRIDER. v. Lidforssii (GEL.) 2 28-35
R. centiformis FRIDER. v. Lidforssii (GEL.) forma 136 28
  R. ambifarius MULL 137 28
{ R. fioniae FRIDER. 3 28
R. fioniae FRIDER. 133 28
R. fioniae FRIDER. 153 28
R. fioniae FRIDER. 12 28-29
{ R. ciliatus LINDEB. 1 28
R. ciliatus LINDEB. 145 +30
  R. » (caesius x Wahlbergii)» f. glaucoformis GUST. 154 28
{ R. pruinosus ARRH. 110 2n=5x=35
R. pruinosus ARRH. 111 35
R. pruinosus ARRH. 143 35
R. pruinosus ARRH. 149 ±35
  R. » caesius x fioniae» 121 35
  Unknown or undetermined 132 2n=4x=28
 

» » »

155

28-29

In addition the chromosome numbers of different other representatives of R. Wahlbergii, nemorosus, gothicus from Småland have been examined. In all cases 2n=4x=28 was obtained. R. balticus ARESCH. from Smâland has 2n== 49, R. fioniae from Ostergotland has 2n=5x=35, R. ambifarius (cultivated in Lund Bot. Garden) has 2n=4x=28, and finally two different forms of R. gothicus from the island of Bornholm have 2n=4x=28 and 2n=5x=35.

2. The East of Skåne.

    No.  
{ R. gothicus FRIDER. p. p. f. microphylla 165 2n=4x=28
R. gothicus FRIDER. p. p. 168 ±28
  R. pruinosus ARRH. v. silvaticus ad int. 157 2n=5x= ±35
  R. ruderalis ARESCH. 185 2n=4x= ±30

II. The West of Sweden (Norway and Denmark).
1. The West of Skane.

    No.  
  R. Wahlbergii ARRH. f. umbrosa 246 2n=5x=35
{ R. centiformis FRIDER. v. Mortensenii FRIDER. et GEL. 243 2n=4x=28
R. centiformis FRIDER. v. Mortensenii FRIDER. et GEL. 239 2n=6x=42
R. centiformis FRIDER. v. Mortensenii FRIDER. et GEL. 179 2n=4x=27-28
  R. »caesius x Mortensenii 188 28
{ R. ciliatus LINDEB. forma 174 28
R. ciliatis LINDEB. v. tiliaceus (ARESCH.) 9 28
R. » ciliatis X eluxatus 175 2n= ±45
  R. (caesius X ciliatus) f. progenerans LIDF. 180 2n=6x=42
  R. eluxatus NEUM. ad subnit, vergens 178 2n=4x= ±28
  R. » caesius x gothicus» f. ruedensis (LIDF.) 245 2n=6x=42
  Unknown 177 42

In 1933 another list of chromosome numbers was published. The chromosome numbers of R. Wahlbergii from Soderâsen, ciliatus and tiliaceus from Kullen were listed as hexaploid, and the numbers of R. » trivultus f. Kullensis» (related to R. ambifarius MULL.?) and of R. eluxatus NEUM.* subnitidus LIDF. as tetraploid and superhexaploid respectively. Since that year two other biotypes have been examined (R. cyclophyllus LINDEB.? and R. permixtus ARESCH., the latter cultivated in Lund Bot. Garden but both originating from Kullen). The chromosome numbers were 6x=42 and 4x=28 respectively.

2. Halland, Bohuslän (Norway).

  R. dissimulans LINDEB. p. p. v. suberectiformis 209 2n=4x=28
  R. hallandicus (GABR.) NEUM. 215 28
{ R. Wahlbergii ARRH. forma 39 28-29
R. Wahlbergii ARRH. 25 2n=5x=35
R. Wahlbergii ARRH. v. partitus GUST 201 35
  R. (caesius x gothicus) f. acutus (LINIJEB.) 210 2n=6x=42
{ R. rosanthus LINDEB. v. leiocarpus LINDEB 43 2n=4x=28
R. rosanthus LINDEB. v. leiocarpus LINDEB 27 2n=7x=49
R. rosanthus LINDEB. v. eriocarpus LINDEB. 41 2n=5x=35
  R. cyclophyllus LINDEB 42 2n=6x=42
  R. Olavii NEUM.. f. 28 ±42
  Unknown 48 2n=4x=28

In 1933 another representative of R. dissimulans (R. nitens), cultivated in Lund Bot. Garden, but originating from Bohuslãn, was examined (6x=42).

3. Denmark.

  No.  
R. dissimulans UNDER. v. selectus 253 2n=6x=42
R. serrulatus LINDEB. f. ditior FRIDER. 266 2n =4x=28
R. Wahlbergii ARRH. v. magnificus FRIDER. 269 2n=5x=35
R. vexatus FRIDER. v. crispus FRIDER. 247 2n =6x=42
R. gothicus FRIDER. 66 2n=4x=28
R. gothicus FRIDER. 72 28
R. gothicus FRIDER. 262 28
R. gothicus FRIDER. 264 28
R. gothicus FRIDER. 268 28
R. centiformis FRIDER.. p. p. 225 28
R. centiformis FRIDER.. p. p. 251 28
R. ambifarius MULL. 249 28
R. ambifarius MULL, f. 184 2n=5x=35
R. fasciculatus MULL.=R. ambifarius MULL.? 274 2n=6x=42
R. fioniae FRIDER. 275 2n=4x=28
R. ferox FRIDER. 258 2n=6x=42
R. ferox FRIDER. 260 ±42
R. Ostenfeldii FRIDER. v. micrander FRIDER. 273 2n=4x=28
R. Ostenfeldii FRIDER.. v. versus Warmingii 223 2n=6x= 42
R. Ostenfeldii FRIDER.. v. versus Warmingii 271 2n=6x= ±44
R. pruinosus ARRH. v. Warmingii (JENS.) 265 2n=5x=35
R. phylloglotta FRIDER. 277 2n=4x=28
R. hystricopsis FRIDER. 267 28
R. » (caesius x Drejeri) » v. aberrans (FRIDER..) 252 28
R. » caesius x Lindebergii» 248 2n=6x=42
»caesius x vestitus» 84 2n=4x=28
R. » caesius x vestitus » 279 28
Unknown or undetermined 280 28
» » » 68 28
» » » 77 28
» » » 276 28
» » » 255 ±30
» » » 256 2n=6x=42
» » » 257 42
» » » 254 ±44

LITERATURE CITED.

  1. BLAKESLEE, A. F. and BUCHHOLZ, J. T. 1929. Pollen-tube growth in crosses between balanced chromosomal types of Datura stramonium. — Genetics 14.
  2. BRAINERD, E. and PEITERSEN, A. K. 1921. Blackberries of New England their genetic status. — Vermont Agr. Exp. Sta. Bull. 218.
  3. DARROW, G. M. 1937. Blackberry and raspberry improvement. — Yearbook of Agriculture, Washington.
  4. FOCKE, W. O. 1914. Species Ruborum. Monographiae Generis Rubi Prodromus. — Stuttgart.
  5. GUSTAFSSON, A. 1930. Kastrierungen und Pseudogamie bei Rubus. — Bot. Not., Lund.
  6. — 1933a. Zur Entstehungsgeschichte des Rubus Bellardii. — Bot Not., Lund.
  7. — 1933b. Chromosomenzahlen in der Gattung Rubus. — Hereditas XVIII.
  8. GUSTAFSSON, A. 1935a. Studies on the mechanism of parthenogenesis. — Hereditas XXI.
  9. —   1935b. The importance of the apomicts to plant-geography. — Bot. Not., Lund.
  10. HAGERUP, O. 1931. Über Polyploidie in Beziehung zur Klima, Okologie und Phylogenie. — Hereditas XVI.
  11. LIDFORSS, B. 1914. Resumé semer Arbeiten liber Rubus.  — Zschr. f. ind. Abst.- u. Vererb.-Iehre 12.
  12. LONGLEY, A. E. 1924. Cytological studies in the genus Rubus. — Amer. Journ. of Bot. 11.
  13. MUNTZING, A. 1936. The evolutionary significance of autopolyploidy. — Hereditas XXI.
  14. ROSANOVA, M. A. 1934. Origin of new forms in the genus Rubus. — Botanical Journal U. S. S. R. 19 (cited from Biological Abstracts 1937, Vol. 11, No. 5).
  15. SCHLÖSSER, L. A. 1936. Befruchtungsschwierigkeiten bei Autopolyploiden und ihre Uberwindung. — Der Züchter 8.
  16. TISCHLER, 6. 1935. Die Bedeutung tier Polyploidie für die Verbreitung der Angiospermen. — Bot. Jahrb. Bd. LXVII.
  17. TURESSON, G. 1929. Zur Natur und Begrenzung der Arteinheiten. — Hereditas XII.

Hurst bibliography