Allozyme polymorphism in isolated populations of the moth Coenophila subrosea ( Lepidoptera : Noctuidae ) from three Central European peat bogs

Allozyme polymorphism was investigated in adult males of the stenotopic peat bog (tyrphobiontic) noctuid moth, Coeno­ phila subrosea Stephens, from three isolated peat bog localities in Austria (Styria) and Czech Republic (South and North Bohemia). Of the eighteen enzyme loci examined, twelve were polymorphic and six monomorphic. Significant deviations of genotype frequen­ cies from Hardy-Weinberg equilibrium were observed at about one third of polymorphic loci within the populations. The average heterozygosities for the populations from three geographically distinct localities ranged from 0.192 to 0.245, and 61% of the loci were polymorphic. The F St mean value of 0.0675 was higher than that found in most other Lepidoptera. The genetic distances based on allozyme heterozygosity ranged from 0.019 to 0.051, with the population from South Bohemia being the most distant. The ge­ netic distances and Fs-r values do not reflect the geographic distances between the populations. Morphometric analysis revealed a dif­ ference between the Austrian Piirgschachen Moor and Bohemian populations. These isolated relict peat bogs are habitat islands inhabited by unique “geographical races” of tyrphobiontic taxa.


INTRODUCTION
Accompanying the fragmentation and destruction of natural landscapes in Central Europe, there has been a si multaneous decline in the number and/or size of many animal and plant populations.Therefore, preserving bio diversity is an important task.Attention has been mainly focused upon relict endangered species of unique natural habitats, e.g.insects, many of them butterflies and moths.The rosy marsh moth, Coenophila subrosea (Stephens, 1829), is known to be closely associated with peat bogs in temperate and southern boreal zones.It is an oligophagous species associated mainly with Myrica gale (in Brit ish Isles) and several other alternative ericaceous bog food plants, e.g.Andromeda polifolia (Tillotson & Spitzer, 1998).Several isolated populations have been re corded in Europe outside Fennoscandia (Spitzer & Novak, 1969;Mikkola & Spitzer, 1983).This boreal moth is a very characteristic tyrphobiontic species.All tyrphobiontic species (tyrphobionts) are stenotopic, usu ally boreal or subarctic faunal components, and obligato rily associated with peat bogs (Warnecke 1926;Peus 1932;Mikkola & Spitzer 1983).The bionomics and Palaearctic distribution of C. subrosea were recently re viewed by Tillotson & Spitzer (1998).To develop an effective conservation strategy for a particular species one needs information not only on its ecological characteris tics but also its genetic variability.There are no popula tion genetic data on the geographical races of C. subrosea in the literature.In this study we examined allozyme polymorphism in, and morphometric data for, three iso lated populations of this species in Central European peat bogs to determine whether these local populations have diverged genetically.The investigated peat bogs are an cient habitat islands of high conservation value.

Study sites
One hundred and thirty adult males were collected at the three sites in Central Europe (Fig. 1): ( 1 (Spitzer et al., 1996).

Electrophoretic analysis
Approximately 40 males were used for the enzyme analyses.They were frozen and stored at -80°C until electrophoresis was done.Before homogenization forewings were removed for mor phometric analysis.Homogenates were made of the thorax plus abdomen of individual moths in 0.1 M Tris -HC1 buffer, pH 7.5, containing 20% glycerol and 0.1% Triton X-100 using a glass homogenizer with teflon pestle.They were then centri fuged at 10,000 g for 20 min at 4°C.All the samples were electrophoresed on vertical 7% polyacrylamide slab gels according to the procedures of Williams & Reisfeld (1964), Davis (1964), or Tris-borate-sulfate system (Allen et ah, 1984).Enzyme ac tivities were visualized using methods of Shaw & Prasad (1970) or Harris & Hopkinson (1976).
The following enzyme systems were screened: adenylate ki nase (Ak, E.C. 2.7.4. Enzymes that gave patterns which were difficult to assign to a genotype, were eliminated from the survey. Allelic variants or allozyme bands for each locus were marked in the order of increasing anodal mobility, a being the allele with slowest mobility.Relative mobilities of the bands are shown in the Table 1.Genotypic frequencies for each polymor phic locus in each sample were tested for fit to the Hardy-Weinberg ratios using y2 goodness-of-fit tests.The observed and  expected heterozygosity (H0 and He, respectively) were calcu lated according to Nei (1973) with the correction for small sam ples (Levene, 1949).The genetic identity 1 and genetic distance D were calculated using Nei's (1978) formula.Subsequent UP-GMA (Sneath & Sokal, 1973)  sions of forewings (length, width, frenulum).Three categories of colour of the forewings was used as a surface character: red, grey and intermediate (see Fig. 4).These data were analysed by one-way ANOVA and Newman-Keuls multiple comparison test using the PRISM graphics programme (GraphPad Software, San Diego, CA, USA).

RESULTS
Twenty three enzyme systems were tested.Twelve en zyme loci: Diaf, Est, a-Gpd, G6pd, Hk, Lap, and Sordh were polymorphic and six other loci: Aat, F6dp, Fum, Itp, Me, Scdh, were monomorphic.Ak, Aid, Pgi, Idh, 6-Pgd were not included in the analysis because of poor or no colour reaction with the stain.Poor resolution of Cat, Est-5, Est-6, Gdh, Mdh and Xdh on gels did not allow a reliable analysis.The Est was the most polymorphic en zyme with six allozymes.On the other hand, Diaf was polymorphic in only one population.
The allozyme frequencies for polymorphic loci are pre sented in Table 1.The observed and expected heterozy gosities for single loci are presented in Table 2 (including monomorphic loci).Significant deviations of genotype frequencies from Hardy-Weinberg expectations were ob served at some polymorphic loci within populations.
Results of the estimates of Nei's (1987) fixation indices are shown in Table 3.Among the populations, analysis of the variance in allele frequency (Fsr) showed significant differentiation (mean FSt= 0.0675) at five of the 18 loci.The fixation index value (Fis) measures the deviation from the proportion of heterozygotes expected for each locus (Table 3).A negative value of Fis which indicates more heterozygotes than expected, was found for Diaf, Est-3, Lap and Sordh, with that for the first two loci being low.For the other loci, the value of F« was positive and the average value was relatively high, 0.187.
A comparison of the populations using all 18 enzyme loci was made to establish genetic identity and genetic distance values.Genetic identity values (Table 4, above diagonal) showed that all populations were genetically differentiated.Moths from North Bohemia and Styria (Austria) were the most similar to each other (I = 0.981), and those that showed the greatest genetic distance were from North Bohemia and the Šumava Mts (South Bohe mia) (D = 0.0505) (Table 4, below diagonal).There was no apparent association between geographic distance and cither genetic distance or mean observed heterozygosity (linear regression, the slopes differ non-significantly from zero -P = 0.929 and P = 0.235, respectively).A dendro gram based on unweighted-pair-groups of the arithmetic averages of genetic distances is presented in Fig. 2.This columns from left to right: Piirgschachcn Moor (Austria), Mrtvÿ revealed that the Sumava Mts population is clearly dis tinct from the other two populations.
Morphometric analysis revealed differences in some of the dimensions of the (brewing in C. subrosea (Table 5, Fig. 3) in the three populations.The population from Sty ria (PQrgschachen Moor) differs from both Bohemian populations in the length and width of the forewing and in the ratio of these dimensions.It is the most distinct popu lation morphologically.The proportion of individuals with greyish coloured forewing is significantly higher in the North and South Bohemian populations, and with a reddish pattern in the Styrian population (Fig. 4).In all three populations a small portion of moths have a reddishgrey fore wing.

DISCUSSION
One of the main objectives of population genetics is to determine the genetic variability in populations and to un-Tabi.k 3. Summary of F-statistics for polymorphic loci in three populations of the moth.Goulson, 1993), for Operophtera brumata (L.) 0.146-0.279(San & Šula, 1993) and for Heliothis zea Boddie 0.270 (Sluss et ah, 1978).The values are also similar to those commonly found in some forest Lepidoptera, 0.22-0.30(Mitter & Futuyma, 1979).However, values as low as 0.083 have been reported for nine Yponomeuta species (Menken, 1982).The average heterozygosity of the populations from South Bohemia (Šumava Mts) and Styria (0.241 and 0.245, respectively) did not differ significantly from that of the North Bohemian population (0.192).The low aver age heterozygosity of the latter population may be a con sequence of it being a tenth of the size of the other populations and, therefore, more subject to the effect of drift and bottlenecks both of which tend to reduce vari ability.Descimon & Napolitano (1993) found that in the endangered sedentary butterfly Parnassius mnemosyne (L.) which often occurs at disjunct colonics, the area of habitat is the most reliable geographical indicator of ge netic diversity.On average, approximately only 6.8% (mean F st = 0.0675, Table 2) of the total variance of allele frequencies was due to genetic differences between the populations of C. subrosea.Thus 93.2% of the genetic diversity (hetero zygosity) is present in each of the populations.The C. subrosea Fs rvalue of 0.0675 is higher than most of the values reported for mobile butterflies and moths; e.g.0.009 in migratory Danaus plexippus (L.) (Eanes & Koehn, 1978), 0.005-0.021 in M. jurtina (L.) (Goulson 1993), and 0.027 in Yponomeuta (Menken, 1982), though values as high as 0.083 for non-migratory ("hostassociated") populations of Zeiraphera diniana Guenée (Emelianov et al., 1995) and 0.154 for Feltia jaculifera (Guenée) (Gooding et al. 1992) have been reported, in more sedentary species, the Fst values are on average usually higher, as in P. mnemosyne populations from Hungary (Meglecz et ah, 1997) and France (Napolitano & Descimon, 1994); 0.064 and 0.135, respectively.

Sample size Fis
The fixation indices (Fn and F|S) computed for each in dividual locus showed persistent heterozygote deficiency at loci Est-1, Est-2, Est-4, a-GPD, G-6-pd and all Hk (both Fn and Fis > 0) and heterozygote excess at loci Diaf, Est-3, Lap and Sordh (both Fit and Fis < 0).Thus, genetic variability at the twelve polymorphic enzyme loci indicates significant deviation from the expected number of heterozygotes.This may be due to selection, subdivi sion of populations or inbreeding.The simultaneous oc currence of a deficit and excess of heterozygotes at vari ous polymorphic loci hint at both substructuring and se lection in the moth populations.The possibility of inbreeding as a cause of prevailing homozygote excess can not be excluded because of small population size, how ever if this were the case one would expect all loci to be affected similarly.
The degree of genetic differentiation between the popu lations of the moth (mean D = 0.038) is comparable to values found in other species.The mean genetic distance between 11 populations of Heliothis virescens (F.) is 0.034 (Sluss & Graham, 1979), for fourteen populations of M. jurtina (Goulson, 1993) it varies between 0 and 0.013, and in conspecific populations of Yponomeuta spe cies it ranges from 0.000 to 0.023, with a mean of 0.006 (Menken, 1981).On the other hand, the genetic distance between two closely related Brenthis species (Nymphalidae) is only 0.043 (Matsuoka et al., 1983).Taking into account that the C. subrosea populations are, in contrast to the above-mentioned examples, geographically and reproductively isolated populations, the mean genetic dis tance is not very high.
There was no apparent association between genetic het erozygosity and the geographical distance between the three isolated populations of C. subrosea.Such associa tion was only observed in the case of wing colour.The absence of an association between genetic heterozygosity in enzyme loci and geographical distance is not excep tional, e.g., some populations of M jurtina more than 150 km apart are genetically identical, while others only 1 km apart are at the upper range of values of D (Goulson, 1993).Meglecz et al. (1997) obtained similar results for populations of P. mnemosyne in North-East Hungary.In C. subrosea, the most likely factor accounting for the di vergence in the populations is separate colonization of the peat bog localities in early Holocene (cf.Spitzer & Novak, 1969;Jankovska, 1980;Spitzer, 1994).The mor- phometric characteristics, colour patterns of the fore wings, and enzyme variation support the hypothesis of a high specificity of each local peat bog population associ ated with unique habitat islands and conform to old ideas (eg.Wamecke, 1926Wamecke, , 1952) ) of specific geographical races ("subspecies") of C. subrosea.Each isolated popu lation is very dependent on its local habitat conservation.
Fig. 3. Relative frequency distributions of the ratio of the width to length of the forewing of populations of C. subrosea.Staré Splavy -open bars, Mrtvý Luh -hatched bars, Purgschachen Moor -solid bars.The difference between popula tions was significant (ANOVA: F2, m = 36.83,P < 0.0001), in that the population at Piirgschachen Moor differed significantly from the other populations at P < 0.05 (Newman-Keuls multi ple comparison test).
T a ble 1. Allele frequencies for polymorphic loci.
Tabi.i: 4. Nei's unbiased estimate of genetic identity (above diagonal) and genetic distance (below diagonal) for three popu lations of C. subrosea.