Apomictic biotypes in Potentilla intermedia and P. norvegica

Material from localities in N. Scania, Sweden, was used in the present investigation. The chromosome number was 2n= 56 in P. intermedia (as previously counted in Polish material) and 2n= 70 in P. norvegica (as in many previous counts). Crossing experiments indicated pseudogamy in both species. Apomictic reproduction has been reported previously in P. intermedia, while P. norvegica has been supposed to be sexual. Embryologic studies gave evidence of obligate apospory in P. intermedia, while in P. norvegica some of the primary EMC:s undergo meiosis. In the latter species, partial sexuality is not quite excluded, although it reproduces mainly by apomixis. Meiosis on the male side is fairly irregular in P. intermedia while on the other hand P. norvegica has a quite regular bivalent formation, in spite of its high degree of polyploidy. P. intermedia is supposed to be of hybrid origin, being the result of crosses between P. argentea and P. norvegica. This remains to be proved experimentally, however, since crossing experiments between P. argentea and P. norvegica have so far failed to give any offspring. Potentilla intermedia L. and P. norvegica L. are supposed to be two fairly closely related species.

P. norvegica has a wide distribution in Europe and Asia. P. intermedia has its main area of distribution in European Russia, from where it has been spreading westwards. Both species are anthropochorous at least in part of their ranges, for instance in Scandinavia. As introduced, they are also found in North America.
Concerning P. intermedia three chromosome numbers have been published. 2n=28 was given by SHIMOTOMAI (1930 a, b; 1935); unfortunately, the origin of his material is unknown. POPOFF (1939) found 2n=42 in collections obtained from the Botanical Gardens of Copenhagen and the Belorussian scientific academy. The number 2n=42 was also found in introduced Canadian material, belonging to var. canescens, by MULLI-CAN (1959); while 2n= 56 was counted in Polish plants by SKALINSKA and CZAPIK (1958). Some of the previous counts cited in MULLIGAN'S paper are erroneously attributed by him to P. intermedia var canescens, and refer instead to the species P. canescens BESSER.
In the surveys on apomixis in the angiosperms by GUSTAFSSON (1947 b) and NYGREN (1954) P. intermedia, but not P. norvegica, is included among species with known or suspected apomixis.
As for P. intermedia, this refers to the only note on the mode of reproduction of this species known to me, by SHIMOTOMAI (1935). The cross P. intermedia (2n= 28) x P. pimpinelloides gave rise to maternal intermedia plants only, while castrated flowers did not set any seeds without pollination. This indicated pseudogamy here, but no embryological investigation on P. intermedia seems to have been carried out.
For P. norvegica, a sexual mode of reproduction was indicated, according to GENTCHEFF (1938). A cross P. norvegica (2n=70)xP. argyrophylln (2n= 63) gave rise to hybrids only, as judged by the morphology of the offspring. However, the chromosome numbers of the F, plants were not determined. Meiosis in P. norvrgica was spoken of as "normal". On the contrary, A. LOVE (1954 a) stated that both P. norvegica and P. monspeliensis "seem to be obligate agamosperms". Neither in these species has the embryology been studied.
Material and methods P. norvegica was collected on the railway bank in the parish of Farstorp, N. Scania. P. intermedia was collected from some localities along highroads in the same parish and in the adjacent one, Vankiva. Although the latter species must be considered as rare in Sweden, it is rather frequent in this area.
Besides, in the crosses listed in Table 1, biotypes of the species P. argenfea, crantzii, tabernaemontani and erecfa were used as staminate parents. The material belonging to the three first-mentioned species was studied by A. MUNTZING (1928A. MUNTZING ( , 1931A. MUNTZING ( , 1958 and A. and G . MUNTZING (1941).
Chromosome numbers were determined on preparates from root tips fixed in the Svalov modification of Navashin-Karpechenko (MUNT-ZING 1933) after refrigeration for at least 12 hours in order to get chromosome contraction. The root tips were embedded in paraffine, cut into sections of 14 p and stained with crystal violet, after bleaching of the cytoplasm in hydrogen peroxide and mordanting in 1% chromic acid.
For meiotic studies, buds were fixed in a 3: 2: 1 mixture of ethyl alcohol, chloroform and acetic acid for 1 hour and then in a 3: I alcohol and acetic acid for 12-24 hours. For staining, alcoholic hydrochloric-acid carmine was used (SNOW 1963). The anthers were squashed in Hoyer's medium (BEEKS 1955).
The percentages of morphologically good pollen were estimated on preparates stained with lactophenolic fuchsine.
Embryological studies were performed on buds fixed in Navashin-Karpechenko after a short prefixation in 1: 3 acetic alcohol. After embedding in paraffine and sectioning, the material was stained with Heidenhain's haematoxylin. Before controlled crossings,. buds were emasculated about one day before opening. The pollination was performed by twirling flowers of the staminate parent around the gynaecia of the pistillate parent.
The distribution of phenolic compounds in leaves was studied by means of thin-layer chromatography according to the Balsgard system (NYBOM 1968;ASKER and FROST 1970).

P. intermedia
The chromosome numbers of two collections from Farstorp and one from Vankiva were found t o be 2n=56, the same as in the Polish ma- terial studied by SKALINSKA and CZAPIK (1958).
No morphological differences could be observed between plants from different localities after cultivation in uniform environment. As, furthermore, the offspring of individual plants was very uniform, the suspicion arose that the plant material belonged to one single apomictic biotype.
To check the mode of reproduction, some crossing experiments were performed in 1964, and were in part repeated in 1965 and 1968 (Table 1). Only interspecific crosses were made, since no other intermedia-types were available.
The crossing result strongly supported the opinion that apomictic reproduction was predominating here. The offspring consisted almost exclusively of maternal plants. Only in the crosses to P. crantzii and P. tabernaemontani, one hybrid plant was obtained in each combination. The hybrids were weak, non-flowering, and in appearance most similar t o their pistillate parent. As their chromosome numbers were found to be 2n= c. 80 and 2n= 77, respectively, they were undoubtedly B 111 hybrids in the sense of RUTIS-HAUSER (1 948), arisen through occasional fertilizations of unreduced egg cells. Castrated flowers gave n o seed-set after isolation, indicating that the intermedia biotype was pseudogamous like other Potentilia apomicts.
Pollen fertility, measured as the percentage of morphologically good pollen, was 64.47, when counted in the summer of 1970 (somewhat lower in earlier counts). On the same occasion, pollen fertility was 82.00,b in P. norvegica and 83.504 in a hexaploid argentea biotype from Lomma.
The rather low pollen fertility in P. intermedia indicated meiotic disturbances, which was also verified by a study of meiosis in the anthers.
The course of meiosis here was in its broad outline the same as that described by Rousr (1965) for the P. anserina complex (Fig. 4-6). Univalents occurred in most cells in first metaphase, as many as 6-7 being sometimes observed.
Laggards were frequent in first anaphase. Probably trivalents and quadrivalents occurred in a few cells in first metaphase.
Embryological studies indicated that P. intermedia is obligately aposporous (Fig. 1-3). The development of the archespore proceeds in the way described by HUNZIKER (1954) and RUTIS-HAUSER (1967). The primary EMC:s arrive a t a pachytene-like stage; n o later meiotic stages were observed. Somatic cells in the chalaza1 region enlarge, become vacuolized and finally give rise t o unrcduced aposporous embryosacs of the usual Polygonurn-type. Often several embryosacs in various stages of development are seen in the same ovule.

P. norvegica
The biotype from Farstorp had the chromosome number 2n=70, which has also been obtained in most previous counts from this species. Also in this species offspring from seeding appeared to be uniform.
In the crosses with P. argentea and P. intermedia, only maternal plants were obtained. If hybrid seeds had arisen, either they did not germinate, or the seedlings died at an early stage. Thus, even here, an apomictic mode of reproduction was indicated. N o B 111 hybrids resulted from the crosses, as in P. intermedia. However, the species P. crantzii and P. tabernaemontani, which gave hybrids when crossed to P. intermedia, were not used as pollinators for P. norvegica (cf. Table I).
The high pollen fertility, as compared to that of P. intermedia, indicated a fairly regular meiosis, in spite of the high degree of polyploidy. Meiotic studies showed that this was the case, and many cells with 35 bivalents were observed (Fig. 7 -9). Univalents were much more scarce than in P. intermedia. If tri-or quadrivalents occurred, they must have been very rare. Unfortunately, the embryologic observations in P. norvegica are still inconclusive. The primary EMS:s sometimes seem t o undergo meiosis, since tetrads were observed in some cases. It is uncertain if reduced embryosacs are formed.
Thus, after pollination by other species, only a parthenogenetic development of unreduced egg cells takes place. A partially sexual mode of reproduction can, however, not be quite excluded, since reduced egg cells might function in other cross combinations. Further crosses, especially with other norv~gica biotypes, as well as continued embryological studies, are needcd t o ascertain the degree and type of apomictic reproduction in P. norvegica. Pseudogamy is indicated   (Fig. 4 6) and P . rlorveKica (Fig. 7-9):. even here, since no seed set occurs after emasculation and isolation.

Meiosis in P . iuti,rn/rdio
The chromatographic pattern of phenolic compounds was studied and compared to that of other Potentilla species by ASKER and FROST (1970). Rather few compounds showed their presence in the chromatograms of P . intermedia and P. norvegica. However, these species showed rather great similarities inter se. Their patterns were also somewhat related to those of P. argentea, but not to those of P. collina, tabernaemontani and crantzii.

Discussion
As stated in the beginning of this paper, P. intermedia has been suspected to represent stabilized hybrids between P . argentea and P. norvegica, in the same way as the series Collinae is thought to have arisen through hybridization between members of Argenteae and Aureae, and P. canescens (by some authors at least) from crosses between Argenteae and Rectae. My primary intention, when beginning this study, was to hybridize P. norvegica (which was then thought to be sexual) with P. argentea in the hope of getting hybrids of intermedia phenotype. Also the chromosome number 2n=56 of the intermedia biotype studied here supported the view that it could have been formed by crosses between P . norvegica (2n=70) and P . argentea (in which 2n=42 is the most frequent number, at any rate outside Scandinavia). The origin of intermedia-biotypes with 2n= 28 and 2n=42 could not, however, be explained in this way.
In this connection the report on sexuality in P. norvegica by GENTCHEFF (1938) should be considered. The Potentilla type assigned by GENTCHEFF to P. norvegica had 2n=70 and, obviously, a sexual reproduction. However, CHRISTOFF and PAPASOVA (1943) had to revise many classifications of Potentilla species in the Botanical Garden of the Agricultural Faculty of Sofia. This material had earlier been studied by POPOFF (1939) who still used erroneous names of several species. Some of the chromosome numbers given by POPOFF are thus open to some doubt, for instance the statement of the numbers 2n= 42 and 2n= 56 in P. norvegica, which have not been met with elsewhere. Furthermore, CHRISTOFF and PAPASOVA found that the "P.
norvegica" from the Botanical Garden of Frankfurt with 2n=70, mentioned by POPOFF, in reality belonged to the species P . rivalis N u n . GENTCHEFF (1938) did not mention the origin of his material, but since at that time he was also working in Sofia, it seems very probable that he used the same plant material as the preceding workers and that his sexual type with 2n= 70 was in reality P. rivalis instead of P. nor vegica .
So far, the attempts to synthesize P. intermedia from crosses between P. norvegica and P. argentea have not been successful. Owing to apomixis in P. norvegica, only maternal plants were obtained from the crosses P. norvegicax argentea. The reciprocal cross (argentea x norvegica), using as pistillate parent a sexual autotetraploid derived from a mainly apomictic diploid (ASKER 1967), failed completely. No seeds were obtained, although the cross was repeated three different times. On the contrary P . intermedia. crossed to the same argentea type, gave rise to some viable and slightly fertile hybrids (ASKER, unpubl.).
The Potentilla norvegica biotype used in this study gave no hybrids following pollination from other species, not even B 111 hybrids which were obtained from P . intermedia in two cases (Table  I). Since the former species has a very extensive distribution area, it is likely that forms will be found with other chromosome numbers than 2n=70, the only one so far ascertained; possibly also sexual types exist somewhere within its range.
Summing up, the origin of P . intermedia remains to be tested experimentally. This species probably consists of apomictic biotypes, which have resulted from crosses between P. argentea and P. norvegica, or perhaps even other members of the series Argenteae and Acephalae (Rivales). To make this clear, material of P. intermedia and its proposed parental species has to be studied from a greater number of localities in different parts of Europe and Asia.