Plant Syst. Evol. 229: 45±57 (2001)

Isozyme evidence of the parental origin and possible fertility of the hybrid Potamogeton ´ ¯uitans Roth J. B. Fant1, C. D. Preston2, and J. A. Barrett1 1 2

Department of Genetics, University of Cambridge, Cambridge, UK Centre for Ecology and Hydrology, Monks Wood, Abbots Ripton, Huntingdon, UK

Received December 1, 2000 Accepted June 20, 2001

Abstract. The hybrid Potamogeton ´ ¯uitans Roth is thought to have resulted from hybridization between P. lucens L. and P. natans L. This hybrid has only been recorded at a few locations in the British Isles. At most of these sites the number of individuals found have been quite low. The one exception to this being a population in the Moors River, Dorset and South Hampshire, England, where a substantially larger population exists. Using seven di€erent enzyme systems, this study provided support for the putative parental origin of this hybrid. In addition, the population in the Moors River was shown to contain numerous multi-locus phenotypes suggesting that these individuals may be partially fertile and may backcross to one of its parental species, or alternatively undergo sel®ng or crossing to produce an F2 generation. Key words: Potamogeton, hybrids, genetic variability, isozymes.

Hybrids are relatively common in the plant kingdom. Although hybridization is a rare event, it has been estimated that hybridization, and subsequent changes in ploidy, is the evolutionary origin of 40±80% of angiosperms (Stebbins 1950, Stace 1975, Rieseberg and Ellstrand 1993). The reasons why hybrids are thought to be more frequent among plants than among animals is due to the reliance

many plants have on external means for mate choice, their potentially greater reproductive longevity, and their potential for vegetative reproduction (Stebbins 1950). The level of fertility found within a hybrid generally depends on the genetic distance between the parental species. Hybrid populations can consist of anything from sterile F1 hybrids, which persist through asexual growth, through to hybrid swarms, containing most of the genetic diversity present in the parental species, fertile F1 hybrids, and backcrosses between them and the ancestral species (Stebbins 1950, Grant 1953, Stace 1975). Backcross progeny with one or both parents are often more frequent than F2s because of lowered pollen fertility in F1 hybrids (Rieseberg et al. 1998). Occasionally, backcrosses with parental species, sel®ng or crossing to other hybrids, can restore fertility, and thereby stabilize the hybrid derivatives and create new species (Stebbins 1950, Stace 1975, Rieseberg 1997). Hybrids are a useful tool in the study of evolutionary processes, including speciation and the exchange of genetic variability. The role of hybridization in the conservation of plant diversity is contentious, being seen as both bene®cial and harmful. In some instances it can lead to an increase in diversity while in

46

J. B. Fant et al.: Isozyme evidence of the parental origin of Potamogeton ´ ¯uitans

others it can result in the extinction of some populations, or species, or at least in the dilution of the `genetic purity' of that species (Ellstrand 1992, Rieseberg and Ellstrand 1993). Even within a species the introduction of alien genes to another population can disrupt local adaptation, a process that Ellstrand (1992) termed `outbreeding depression'; a condition that would be exacerbated with interspeci®c crosses. However, as some species reach critically low numbers, the lack of potential mates may result in hybridization being one of the means by which a population can survive. Potamogeton is one of two genera in the Potamogetonaceae family, the other being the monotypic Groenlandia. It is found in most areas of the world, except arid and polar regions, making it one of the more important aquatic vascular plant genera (Preston 1995). The chromosome numbers of Potamogeton species range from 2n ˆ 14 to 2n ˆ 104, suggesting that both polyploidy and aneuploidy have been important events in the evolution of this genus (Les 1983, Les and Philbrick 1993, Hollingsworth et al. 1998a). Potamogeton species vary considerably in their ecological and morphological characteristics. All pondweed species are potentially able to reproduce both sexually and asexually, although little is actually known about the importance of reproduction by seed within Potamogeton species and few examples of seedlings have been reported from the wild (Preston 1995). Previous studies of Potamogeton populations have shown that genetic variation within these plants, like other hydrophilous species, is predominantly greatest between populations and limited within them (McMillan and Phillips 1981, Barrett and Shore 1990, Hollingsworth et al. 1995a, Hollingsworth et al. 1998b), suggesting that clonal growth is playing an important role in colonising sites. Many of the populations investigated in previous studies were found to be monomorphic, suggesting that these populations arose from a single founder (Hofstra et al. 1995, Hollingsworth et al. 1996a, Hollingsworth et al. 1996b, Hollingsworth et al.

1998b). Hollingsworth (1996) identi®ed clonal growth, low levels of sexual reproduction and geographical isolation as the main reasons for the lack of genetic variation within populations. An important characteristic of Potamogeton, as is the case with many aquatic species, is the prevalence of interspeci®c hybrids. In a study of 177 aquatic genera, hybridization has been reported in 26%, once single-species genera are excluded (Les and Philbrick 1993). Within the family Potamogetonaceae alone there are at least 50 hybrids recognized worldwide (Wiegleb and Kaplan 1998). Much of the evidence for the existence of hybrids has been inferred from their sterility and morphological intermediacy between the putative parents. This is despite the observation that most aquatic species show large morphological variation between di€erent environments, and that some hybrids are found at sites where neither parent exists. The common occurrence of hybrids in aquatic species is undoubtedly aided by the high incidence of vegetative propagation, which can help a hybrid invade a site and persist. Most hybrids are believed to be derived from interspeci®c crosses although it has been proposed that some of these individuals identi®ed as hybrids may be the result of autopolyploidy (Les and Philbrick 1993). Although most of the evidence for the existence and origin of hybrids in the genus Potamogeton comes from morphological studies, these characters tend to be unreliable for population studies due to varying heritability, dominance, epistasis and pleiotropy; hence biochemical and genetic markers are often preferred. Allozymes, which are co-dominant, have simple Mendelian inheritance and are direct gene products, have been used successfully in determining the origin of hybrids in many genera, including Potamogeton (Gallez and Gottlieb 1982, Hollingsworth et al. 1995b, Hollingsworth et al. 1996b). The presence of variation within hybrid populations is likely to be due to independent hybridization events followed by vegetative reproduction; segregation from intercrosses between independently produced hybrids, and backcrosses of hybrids

J. B. Fant et al.: Isozyme evidence of the parental origin of Potamogeton ´ ¯uitans

to one or other parent. As most hybrids are sterile, independent hybridization events are the more likely cause. Variation in this instance originates from genotypic di€erence in the parental species involved in the original crosses. With allozyme banding patterns, it would be expected that the F1 hybrids will consist of additive combinations of the parental bands. If on the other hand the variation is due to sexual reproduction, then segregation of the hybrid combinations will occur. In these instances the random assortment of genes that occurs during segregation may result in the loss of some characteristic parental bands, which may help to distinguish sexual reproduction from independent hybridization. However the absence of expected bands in the hybrid may also be due to insucient sampling of parental populations, and hence missing some multi-locus phenotypes that may have been involved in the original cross. On the basis of morphology, the hybrid P. ´ ¯uitans Roth is believed to result from the crossing of P. lucens L. and P. natans L. It has been found at only a few sites in the British Isles, although in the Moors River, in South Hampshire and Dorset it is, locally, quite common. At the other sites in Cambridgeshire, Huntingdonshire and Norfolk the numbers are much lower. P. ´ ¯uitans was ®rst collected from the Moors River in 1870, although it was initially mistaken for its putative parent P. lucens (Dandy and Taylor 1939). Plants in this population have been found to produce well-formed fruit (Preston 1995). This might suggest partial fertility, although it is not certain whether these fruit contained viable seeds. Work carried out on this population by Hollingsworth (1995) was inconclusive about the amount of genetic variability present in this population. He identi®ed nine di€erent variants of the hybrid, suggesting that this hybrid arose on nine separate occasions, or, possibly that these populations are partially fertile and are segregating. However, this result is in part based on a single enzyme system, SKD, for which he was able to identify six allozyme variants. He also investigated the potential

47

parentage of P. ´ ¯uitans as being a hybrid between P. lucens and P. natans. Although most bands could be attributed to one or both parents, two enzyme systems, SKD and PGM, both produced novel bands. Another discrepancy was the presence of a common band in both P. lucens and P. natans that was not present in the hybrid. This was explained as either too few individuals from the parental species being sampled or as evidence for backcrossing, or sel®ng, of the hybrid (Hollingsworth 1995, Hollingsworth et al. 1995b). The aim of this study is to gather biochemical evidence that P. ´ ¯uitans arose from the hybridization of P. lucens and P. natans, and to determine how much genetic diversity occurs within the Moors River populations using seven isozymes systems. If the hybrid is sterile it is expected that any variation found will most likely have originated from genotypic di€erence in the parental species involved and therefore the banding pattern of the resulting hybrids will still consist of all parental bands plus any intermediate bands, i.e. a characteristic F1 hybrid banding pattern of dimeric and tetrameric enzymes. If there is segregation of the hybrid's banding pattern then it would imply that the variation is due to sexual reproduction. In these situations the banding pattern of the hybrids will be determined by the type of the cross (i.e. intercross (sel®ng) versus backcross). Materials and methods Plant material. Samples of P. ´ ¯uitans were collected from three river or ditch systems in England (Table 1). The largest population occurred in the Moors River, while the remaining populations were found in East Anglia, in ditches at Chear Fen, Cambridgeshire and Woodwalton Fen, Huntingdonshire. There are records of P. ´ ¯uitans occurring at Limpenhoe Marshes in Norfolk in 1989, but no individuals were found when this site was visited in 1997. The number of plants at both East Anglian sites was small, with only one individual being located at Woodwalton and two at Chear Fen. Samples of the parent P. lucens were only collected from Chear Fen, as this was the only site at which

Chear Fen, v.c. 29 Chear Fen, v.c. 29 Chear Fen, v.c. 29 Chear Fen, v.c. 29 Woodwalton Fen, v.c. 31 Chear Fen, v.c. 29 Woodwalton Fen, v.c. 31

P. ´ ¯uitans

P. natans

S. of Hurn, v.c. 11 St Leonards Peat South, v.c. 9/11 St Leonards Peat South, v.c. 9/11 Hurn Forest, v.c. 11

P. natans

P. lucens

S of Hurn, v.c. 11 St Leonards Peat South, v.c. 9/11 St Leonards Peat South, v.c. 9/11 Hurn Forest, v.c. 11

P. ´ ¯uitans

Cambridgeshire (v.c. 29) & Huntingdonshire (v.c. 31)

Sopley Bridge (River Avon), v.c. 11 North of Christchurch (River Avon), v.c. 11 River Frome, v.c. 9

P. lucens

Dorset (v.c. 9) & South Hampshire (v.c. 11) (Moors River unless stated)

Location & sites

Species

Locality

TL(52) 49-70 TL(52) 23-84-

TL(52) 49-70TL(52) 23-84-

TL(52) 49-71TL(52) 49-70TL(52) 48-70

SZ(40) 12-96SU(41) 09-01SU(41) 09-00SU(41) 10-00-

SZ(40) 12-96SU(41) 09-01SU(41) 09-00SU(41) 10-00-

SZ(40) 14-97SZ(40) 15-94SY(30) 84-87-

Grid ref.

4 3 (7)

2 1 (3)

5 3 9 (16)

2 5 2 2 (11)

6 6 4 2 (18)

3 1 2 (6)

No. of ramets

1 1 (2)

1 1 (2)

1 1 2 (2)

1 3 1 1 (5)

4 3 3 2 (6)

1 1 1 (3)

No. of multi-locus phenotypes

Table 1. List of sites from which the P. ´ ¯uitans and its putative parents were collected, with numbers of ramets sampled and di€erent multilocus phenotypes identi®ed at each site

48 J. B. Fant et al.: Isozyme evidence of the parental origin of Potamogeton ´ ¯uitans

J. B. Fant et al.: Isozyme evidence of the parental origin of Potamogeton ´ ¯uitans both the hybrid and this parent were growing together. As no P. lucens has yet been identi®ed in the Moors River, plants were collected from the neighbouring river systems, the River Avon and the River Frome. P. natans was found at all of the sites from which P. ´ ¯uitans could be collected. As both Potamogeton species and the hybrid reproduce vegetatively and tend to form a relatively large plant mass, samples were taken at least 25 metres apart, unless ramets could be clearly distinguished. However, in most cases, two samples were taken from each collection point to test whether clumps of plants were of the same genotype (i.e. single ramet). All samples were placed in plastic bags ®lled with water and stored in the dark until they could be taken back to the laboratory. Electrophoresis. Isozyme extractions were made from the non-¯owering, growing tips of Potamogeton. Protein extractions were made within 3 days of collection. Extractions were made by vortexing 2 g of leaf tissue in an OakridgeÒ tube, containing a small amount of type A-5 Alumina, 2 ´ 10 mm hard-porcelain grinding balls and liquid nitrogen. Once a ®ne powder was produced, 2 ml of extraction solution (0.01 M Tris, pH 7.2, 0.01 M KCl, 0.01 M MgClá6H20, 0.001 M EDTA, 4% PVP, 1% Triton X-100, 10% v/v glycerol, 0.006 M DTT, 1% v/v mercaptoethanol, 0.2% w/v BSA, 10% DMSO) was added and the tubes were placed on ice for 1±3 hours. The resulting solutions were transferred to 1.5 ml Eppendorf tubes and spun for 10 minutes to remove precipitates. Then 10 aliquots of 200 ll were collected from the supernatant and stored at )70 oC for up to 2 years. Seven enzyme systems were used to investigate the population diversity and parentage of P. ´ ¯uitans, for the three locations ± AAT (EC 2.6.1.1), ADH (EC 1.1.1.1), GDH (EC 1.4.1.2), IDH (EC 1.1.1.42), ME (EC 1.1.1.40), PGD (EC 1.1.1.44) and SKD (EC 1.1.1.25). A discontinuous polyacrylamide system was used to resolve the isozyme bands as described by Hames (1984). The gels were cast using Hoefer SE 600 series vertical slab gel kits. The pH, acrylamide percentage and composition of the resolving and stacking gel solutions varied according to the enzymes being studied. Enzymes systems AAT, GDH, IDH and PGD, were run on a high pH system with resolving gel of 9% for GDH, 11% for AAT & PGD and 15% for IDH, and a Tris-glycine bu€er electrophoretic bu€er at pH 8.3. While enzymes systems ADH, ME and SKD where run

49

on neutral pH gel system at 9% for ADH & ME and 15% for SKD, in a Tris-borate electrophoretic bu€er (pH 7.5 or 7.0). The wells were washed with electrode bu€er prior to loading the 20 ll of the protein extracts. The gels were run at a constant current of 10 mA per gel, and 240 V, for 18 hours, or until the yellowish protein front had run to the bottom of the gel. A heat exchanger was used to keep the electrode bu€er cool. Each run consisted of 2 assemblies, containing a total of eight individual gels, allowing for eight di€erent stains to be used. Gels were covered with 50 ml of a stain solution speci®c to the enzyme being investigated and placed on a rotator in the dark. Stains were made up according to Wendel and Weeden 1989. A photographic record of all gels was made, and scored for the presence and absence of bands.

Results The plants of P. ´ ¯uitans collected from the Moors River were morphologically quite distinct in appearance from those found at the two East Anglian sites (Fant 1999). The Moors River is slow ¯owing and produces largish clumps of plants with long stems with a greater proportion of submerged leaves and only a few sub-coriaceous leaves at the tips of the stem; apart from these sub-coriaceous leaves it looked very much like its putative parent P. lucens. The East Anglian material has both submerged and ¯oating leaves more similar to those of the other putative parent, P. natans. The di€erence between plants from Chear Fen and Woodwalton Fen was less pronounced than their di€erences from the Moors River material, but nonetheless was also distinctive, with the clump at Woodwalton producing submerged leaves which were much more phyllode±like than those of the Chear Fen material. Parental origins of the P. ´ ¯uitans Of the seven enzyme systems used, only for IDH was there no variation between the two putative parents and the hybrid (Fig. 1; Table 1). As most of the enzymes systems used in this study are polymeric it is dicult to say with any certainty if the banding pattern seen

50

J. B. Fant et al.: Isozyme evidence of the parental origin of Potamogeton ´ ¯uitans

J. B. Fant et al.: Isozyme evidence of the parental origin of Potamogeton ´ ¯uitans

51

Fig. 1. Diagrammatic representation of the di€erent banding patterns found in the two parental species (P. lucens and P. natans) and its putative hybrid (P. ´ ¯uitans). M.R. Moors River, R.F. River Frome, R.A. River Avon, C.F. Chear Fen, W.W. Woodwalton

is purely of parental origin, or resulting from segregation or backcrossing. The exception is SKD, which is the only monomeric system used in this study. The banding pattern of SKD was complex, being comprised of numerous bands of di€erent intensity. This made it dicult to distinguish individual bands, and for this reason a more conservative estimate of the number of phenotypes was taken. Nonetheless, four clear phenotypes could be identi-

®ed in the Moors River. Unfortunately, Woodwalton material did not stain well, so no estimate of variation could be made for East Anglia. As might be expected of a F1 hybrid, the Chear Fen P. ´ ¯uitans pattern contained all the bands found in the two parental patterns, although they di€ered in the intensity of staining. However with the Moors River population three phenotypes (D, F & G) could be explained by the crossing of one of

52

J. B. Fant et al.: Isozyme evidence of the parental origin of Potamogeton ´ ¯uitans

the P. lucens with either phenotype J or L of P. natans, while the fourth phenotype (E) seems to be lacking a band found in most individuals examined, with the exception of a single P. natans. This could be that we did not ®nd the original P. lucens parent that was involved in this cross or that the population is the result of segregation. For the other enzyme systems all the bands of the hybrids were either present in one of the parental banding patterns, or could be explained as an intermediate band formed through the combining of parental bands. AAT di€ered from other enzymes in that in some cases banding patterns in the hybrids were almost identical to that of one of the parental species. For this enzyme, hybrid material from the Moors River tended to resemble P. lucens, except for one band in genotype F that was identi®ed as belonging to P. natans, while at the East Anglian sites more of the bands could be attributed to P. natans. The banding pattern of the hybrid for ME did not show any major di€erence between hybrids and regions. ME is a tetramer; consequently the genetics of the inheritance of each individual band can be dicult to decipher due to the large array of possible bands that can result from a single cross. Hence it is not surprising that the banding patterns of the hybrid are not directly related to the parental banding pattern. PGD was the one enzyme that re¯ected regional di€erences best, with P. natans in particular showing marked di€erences between the Moors River and East Anglia, which was also expressed in the banding pattern of the hybrid. In the Moors River the hybrid produced a range of band sizes similar to that of P. lucens, however it had more intermediate bands. Banding pattern E of PGD had the full complement of bands from both parents, while both D & F lacked two bands, both from the P. natans in the latter case, and one from each parent in the former. The banding patterns of the East Anglian hybrids seemed to contain no upper band that appeared in both parental species, although this band is very faint in P. natans. The lack of this band could possibly

be a consequence of it being too faint to detect, given that they were already faint in P. natans and one of the P. lucens genotypes, or alternatively it may be due to too small a sampling size for the parental species. Diversity A comparison of the banding patterns of P. ´ ¯uitans with those of its putative parents, P. lucens and P. natans, showed that all of its bands could be attributed to one or other of the parents, or a combination of the two (Fig. 1). In the parental species, 3 multi-locus phenotypes of P. lucens and 5 multi-locus phenotypes of P. natans were identi®ed in the populations at Dorset and South Hampshire, while two di€erent multi-locus phenotypes of each species occurred in East Anglia (Fig. 1). In P. lucens variability between sites seemed greater than within sites, while P. natans populations were comprised of a number of multi-locus phenotypes with no obvious pattern of distribution throughout the river systems. Eight di€erent multi-locus phenotypes of P. ´ ¯uitans could be identi®ed over the six polymorphic systems. The hybrid material from East Anglia di€ered from that of Dorset and South Hampshire in ®ve of the seven enzymes used: AAT, ADH, GDH, PGD and SKD (Fig. 1). Two of these phenotypes were restricted to East Anglia, one from Chear Fen and the other from Woodwalton Fen, while the remaining six occurred in the Moors River. The di€erences between the two East Anglia sites, Chear Fen and Woodwalton Fen, were less clear, with di€erences only being detected for AAT, PGD and ADH. The distribution of multi-locus phenotypes of P. ´ ¯uitans in the Moors River proved to be very variable, with most being scattered throughout the river, much like P. natans. The populations of the hybrid at the East Anglia sites had no withinpopulation variation. In the Moors River hybrid population, one ramet of the hybrid appeared to be unique, producing distinctive banding patterns for AAT (D), ADH (D), PGD (D) and SKD

J. B. Fant et al.: Isozyme evidence of the parental origin of Potamogeton ´ ¯uitans

53

Table 2. Phenotypes found in P. ´ ¯uitans and its putative parental species Species

Location

No. of ramets

AAT

ADH

GDH

IDH

ME

PGD

SKD

P. lucens

River Avon River Avon Frome River Chear Fen Chear Fen

3 1 2 13 3

A B A C C

A A A B C

B B A C C

A A A A A

B C A D D

A A A B C

B B A C C

P. ´ ¯uitans

Moors River Moors River Moors River Moors River Moors River Moors River Chear Fen Woodwalton

1 4 5 2 4 2 2 1

D E F E F F G H

D E F E F F G H

C C C C C C D D

A A A A A A A A

E F G G E G E E

D E E F F F G H

E D F G G G H H

P. natans

Moors River Moors River Moors River Moors River Moors River Chear Fen Woodwalton

1 3 3 2 2 4 3

I J J K ±a K J

J K J I J J J

D D D D D D D

A A A A A A A

H I J J K H L

I I I I I J K

I J J K L L J

a

Non-resolved banding pattern

(E). In the case of AAT the pattern for this ramet was exactly the same as that of P. lucens from the River Avon, but the pattern for SKD and ADH was closer to that of P. natans than P. lucens. In addition, the pattern it produced for PGD was unlike either parental species or any of the other hybrids in that it produced multiple bands that were clustered into three regions. This re¯ected the banding pattern of P. lucens, which produced more intense bands in these same regions, but only as single bands. This ramet did, however, share its banding pattern for ME with four other ramets collected in the Moors River, as was the case with GDH and IDH. Although P. lucens collected from the River Frome and River Avon had distinctive banding patterns, it was nonetheless not possible to determine with any great certainty which of these particular forms was more likely to be the parental source of the hybrid. In the case of AAT, ME, PGD and SKD the

di€erences in banding pattern were not great enough to determine which parent was the source of speci®c bands, especially as the di€erences were mainly due to relative intensity of staining. It is, however, likely that the P. lucens parent of the Moors River hybrids was probably more similar to the phenotype found in the Frome River than to the River Avon population. Discussion This study supports the current theory that P. ´ ¯uitans arose as a hybrid between P. natans and P. lucens. For all enzymes used the bands of the hybrid could be explained as corresponding directly to one or other parental species or as an heteromer between the two parental forms. P. ´ ¯uitans is relatively rare, being found at only four sites around Britain, despite both of its parents being quite common British species which can be frequently found

54

J. B. Fant et al.: Isozyme evidence of the parental origin of Potamogeton ´ ¯uitans

growing together at the same sites. This might suggest that genetically these species are quite far apart, preventing the formation of hybrids, although Moors River material tends to suggest that they are still suciently similar to produce a partially fertile hybrid. The other possible explanation for its relative rarity is that the hybrid is less ®t than its parents are and is being out-competed. Of the four sites at which this hybrid has been recorded, three of them do not contain the parent P. lucens. This may suggest that the reason we do not see more P. ´ ¯uitans is that they are being outcompeted by P. lucens. In this allozyme study, the analysis was complicated by the fact that many of the enzymes are polymeric, both P. lucens and P. natans are polyploid in origin (and hence have multiple copies of some enzymes) and we have no known F1 to compare to the samples studied. Segregation and backcrossing can be dicult to separate, with many polymeric isozymes. However, with monomeric isozyme systems one allele gives one band, so an F1 hybrid would express single copies of parental bands (Wendel and Weeden 1989). With SKD (a monomeric enzyme) Hollingsworth et al. (1995b) found a band in P. ´ ¯uitans that was not present in either parent. In this present study all the bands for SKD could be attributed to one or both parental species. It is possible that the appearance of the `additional' band interpreted by Hollingsworth et al. could be a result of starch gels not suciently distinguishing the individual bands from the triplet of bands seen in the middle region of the PAGE gels. One other interesting di€erence that Hollingsworth et al. (1995b) noted was the presence of a faster band in all of the P. natans, but none in the hybrid. This is not what is expected in an F1 hybrid with a monomeric enzyme. In this study all P. ´ ¯uitans possessed this faster band although it stained much more weakly in some genotypes. However, in this study one variant of P. natans was found which did lack the faster travelling band and which therefore could be the potential parent of these hybrids.

The two East Anglian populations of P. ´ ¯uitans were small and consisted of only a few plants. The scarcity of plants in Woodwalton Fen might suggest that these plants are not well suited to this site. The population at Chear Fen is relatively new, not being recorded in the vicinity since 1897. The lack of previous records may be due to the population being overlooked, especially as the plants tended to be obvious only in the early part of the season, or possibly because it is a new arrival. The former seems the likely explanation, as later in the season, even with good records of its location and numerous trips to the site, these plants could not be easily distinguished from P. natans. This population di€ers quite considerably from the Dorset populations, where the hybrid is distinctive in having an almost P. lucens-like appearance, but this could be a result of the environment in which they are growing. The degree of morphological similarity between the two East Anglian sites, as compared to the Moors River material, was re¯ected in the allozyme patterns. Woodwalton Fen and Chear Fen shared the same banding pattern for four enzymes, but di€ered in the remaining three: ADH, PGD and SKD. Even for these three enzymes, the Woodwalton and Chear Fen plants were more similar to each other than to P. ´ ¯uitans from the Moors River. This is most likely a result of the parental individuals that were involved in the original crosses being more genetically similar within these East Anglian populations. At the East Anglian sites only one ramet of P. ´ ¯uitans was found at Woodwalton Fen and two at Chear Fen so it was not possible to ®nd any genotypic di€erence within these sites. Therefore we can only say that there was only a single hybridization event at each site. In the Moors River population, six di€erent multi-locus phenotypes were identi®ed. This would suggest either that the hybrid in this river has arisen on numerous occasions, or that there is some partial fertility that has allowed the hybrid to reproduce sexually, either through backcrossing or self-fertiliza-

J. B. Fant et al.: Isozyme evidence of the parental origin of Potamogeton ´ ¯uitans

tion. These results are supported by previous work on P. ´ ¯uitans from this site that also identi®ed a high level of allozyme diversity (Hollingsworth et al. 1995b). In general the banding pattern for the four enzyme-systems common to Hollingsworth et al. (1995b) and this study were comparable, producing similar patterns and similar levels of diversity. As hybrids are generally thought to be sterile, the most likely explanation for the presence of these di€erent multi-locus phenotypes would be multiple hybridization events. For hybrids to arise this frequently, implies that the barriers to hybridization between P. lucens and P. natans are small. This seems unlikely, as both these species are relatively common in England and frequently occur at the same sites, so the frequency of P. ´ ¯uitans in Britain would be expected to have been much greater than it currently is. Multiple origins would also imply that there was a high frequency of contact between parental species. This is unlikely to have happened in recent years as no P. lucens was found in the Moors River and there are no records of it ever having been found at this site, although it does occur in the neighbouring river systems, the Rivers Frome and Avon. Hence any new hybrids could only arise through P. lucens pollen reaching the P. natans in the Moors River, and also outcompeting maternal pollen. As interspeci®c pollen competition is one of the major mechanisms that prevent hybridization, this seems unlikely (Rieseberg and Carney 1998). Alternatively the crosses could occur elsewhere, and the seeds arrive at this site via some regular dispersal method, such as water currents. However, as the Moors River is not a branch of any other river, and as P. lucens does not occur upstream from the sites sampled, it seems improbable that new hybrid seed would constantly end up at this one site. Another possibility is that the genotypes of P. ´ ¯uitans in the Moors River date from multiple hybridization events in the past, at a time when P. lucens may have been present in the river. The possibility that this population is a relic hybrid swarm would explain why the Moors

55

River population does resemble P. lucens and their apparent increase in fertility. Given that the conditions for multiple origin seem unfavourable, the only alternative source of variation in these populations is that this hybrid is capable of sexual reproduction. This is supported by the production of what seems to be well-formed fruit, which suggests that P. ´ ¯uitans may be partially fertile (Preston 1995). The most likely method of sexual reproduction in hybrids is backcrossing, because the semi-fertile pollen of hybrids is usually out-competed by fertile pollen from parental species, although this can depend to some degree on the density of parental species (Stebbins 1950). As P. natans is the only parental species in the area, it makes it the most likely species with which the hybrid is backcrossing. However, the morphology and some of the allozymes systems do not match this explanation, in particular GDH and AAT where the P. ´ ¯uitans banding pattern resembles more closely that of P. lucens. The possibility that pollen from P. lucens is fertilizing these hybrids seems equally unlikely as the nearest source is the River Avon, some 2±3 kilometres away. Hence it seems that the segregation of F2 hybrids might be the more likely explanation. This sort of event would result in the production of F2 hybrids which would be expected to produce a greater degree of variability in morphological characters, overlapping the distribution of parental characters, than was found in this population. However the environment can heavily in¯uence the morphological characteristics of most Potamogeton spp. so they are not a reliable measure of genotypic di€erence (McMillan and Phillips 1981, Preston 1995). The lack of obvious morphological segregation in the Moors River populations may be a result of the environment playing a larger role on the phenotype than genotypic di€erences. Considering in a stable population that the success of seedling establishment for a freely fruiting Potamogeton species appears to be low, it seems less likely that a rarely fruiting F1 hybrid could produce an F2 generation.

56

J. B. Fant et al.: Isozyme evidence of the parental origin of Potamogeton ´ ¯uitans

However in this study it was found that the number of multi-locus genotypes of P. ´ ¯uitans was equivalent to that of P. natans, which would suggest that in the Moors River this hybrid is as successful at regenerating from seed as its parent. This might also suggest that the hybrid has a relatively high level of fertility. Buerkle et al. (2000) were able to produce models that showed that speciation could occur in homoploid hybrids as a result of recombination through backcrossing or inbreeding between F1 individuals. In the Moors River there was one individual that produced a banding pattern that suggested that it resulted from a backcross to P. lucens. In this individual the banding patterns of AAT, PGD and SKD were more similar to those of P. lucens than compared to other hybrids. The possibility that this individual was a P. lucens, rather than a hybrid, was discounted as the banding patterns for other enzymes were consistent with those of other P. ´ ¯uitans, and the plant was morphologically similar to other hybrids. This individual also occurs in a region of the Moors River close to the river Avon where P. lucens is found. If this individual was the result of a backcrossing event it may explain the apparent stabilisation and restoration of fertility within this population (Stebbins 1950, Stace 1975, Buerkle et al. 2000). It would seem clear from this study that P. ´ ¯uitans is the result of a cross between the two species P. natans and P. lucens. This is a relatively rare hybrid arising at only four locations in the British Isles, and most populations are small, consisting of single ramet, resulting from a single hybridization event. However, the Moors River population is larger and the isozyme evidence suggests that the plants there are at least partially fertile, and reproducing sexually. However, this study was unable to ascertain if this was a result of backcrossing, sel®ng or both. References Barrett S. C. H., Shore J. S. (1990) Isoenzyme variation in colonising plants. In: Soltis D. E.,

Soltis P. S. (eds.) Isozymes in Plant Biology. Chapman and Hall, London, pp. 106±126. Buerkle C. A., Morris R. J., Asmussen M. A., Rieseberg L. H. (2000) The likelihood of homoploid hybrid speciation. Heredity 84: 441±451. Dandy J. E., Taylor G. (1939) Studies of British Potamogetons. J. Bot., London 77: 342±343. Ellstrand N. C. (1992) Gene ¯ow by pollen: implications for plant conservation genetics. Oikos 63: 77±86. Fant J. B. (1999) ``The conservation and maintenance of genetic diversity both in situ and ex situ''. PhD Thesis, Department of Genetics, University of Cambridge. Gallez G. P., Gottlieb L. D. (1982) Genetic evidence for the hybrid origin of the diploid plant Stephanomeria diegensis. Evolution 36(6): 1158±1167. Grant V. (1953) The role of hybridisation in the evolution of the leafy stemmed Gilias. Evolution 7: 51±64. Hames B. D. (1984) An introduction to Polyacrylamide Gel Electrophoresis. In: Hames B. D., Rickwood D. (eds.) Gel Electrophoresis of Proteins. Oxford University Press, Oxford, pp. 1±86. Hofstra D. E., Adam K. D., Clayton J. S. (1995) Isozyme variation in New Zealand populations of Myriophyllum and Potamogeton species. Aquatic Bot. 52: 121±131. Hollingsworth P. M. (1995) Population genetics of some species of Potamogeton L. PhD Thesis, Department of Botany, The University of Leicester. Hollingsworth P. M., Gornall R. J., Preston C. D. (1995a) Genetic variability in British populations of Potamogeton coloratus (Potamogetonaceae). Plant Syst. Evol. 197: 71±85. Hollingsworth P. M., Preston C. D., Gornall R. J. (1995b) Isozyme evidence for the hybridisation between Potamogeton natans and P. nodosus (Potamogetonaceae) in Britain. Bot. J. Linn. Soc. 117: 59±69. Hollingsworth P. M., Preston C. D., Gornall R. J. (1996a) Genetic variability in two hydrophilous species of Potamogeton, P. pectinatus and P. ®liformis (Potamogetonaceae). Plant Syst. Evol. 202: 233±254. Hollingsworth P. M., Preston C. D., Gornall R. J. (1996b) Isozyme evidence for the parentage and multiple origins of Potamogeton ´ suecicus (P.

J. B. Fant et al.: Isozyme evidence of the parental origin of Potamogeton ´ ¯uitans pectinatus ´ P. ®liformis, Potamogetonaceae). Plant Syst. Evol. 202: 219±232. Hollingsworth P. M., Preston C. D., Gornall R. J. (1998a) Euploid and aneuploid evolution in Potamogeton (Potamogetonaceae): a factual basis for interpretation. Aquatic Bot. 60: 337±358. Hollingsworth P. M., Preston C. D., Gornall R. J. (1998b) Lack of detectable isozyme variability in British populations of Potamogeton epihydrus (Potamogetonaceae). Aquatic Bot. 60: 433±437. Les D. H. (1983) Taxonomic implications of aneuploidy and polyploidy in Potamogeton (Potamogetonaceae). Rhodora 85: 301±323. Les D. H., Philbrick C. T. (1993) Studies of hybridisation and chromosome number variation in aquatic angiosperms: evolutionary consequences. Aquatic Bot. 44: 181±228. McMillan C., Phillips R. C. (1981) Morphological variation and isozymes of North American Phyllospadix (Potamogetonaceae). Canad. J. Bot. 59: 1494±1500. Preston C. D. (1995) Pondweeds of Great Britain and Ireland. B.S.B.I., London. Rieseberg L. H. (1997) Hybrid Origin of Plant Species. Ann. Rev. Ecol. Syst. 28: 359±389. Rieseberg L. H., Baird S. J. E., Desrochers A. M. (1998) Patterns of mating in wild sun¯ower hybrid zones. Evolution 52(3): 712±726.

57

Rieseberg L. H., Carney S. E. (1998) Plant hybridization (Tansley Review No. 102). New Phytol. 140(4): 599±624. Rieseberg L. H., Ellstrand N. C. (1993) What can molecular and morphological markers tell us about plant hybridisation? Crit. Rev. Pl. Sci. 12(3): 213±241. Stace C. A. (1975) Hybridization: and the Flora of the British Isles. Academic Press Inc., London. Stebbins G. L. J. (1950) Variation and Evolution in Plants. Columbia University Press, Columbia. Wendel J. F., Weeden N. F. (1989) Visualization and Interpretation of Plant Isozymes. In: Soltis D. E., Soltis P. S. (eds.) Isozymes in Plant Biology. Chapman and Hall, London, pp. 5±45. Wiegleb G., Kaplan Z. (1998) An account of the species of Potamogeton L. (Potamogetonaceae). Folia Geobot. 33: 21±316.

Addresses of the authors: J. B. Fant (e-mail: [email protected]), J. A. Barrett, Department of Genetics, University of Cambridge, Downing Site, Cambridge CB2 3EH, UK. C. D. Preston, Centre for Ecology and Hydrology, Monks Wood, Abbots Ripton, Huntingdon PE 28 2LS, UK.