The intertidal snail Thais clavigera (Küster, 1858) has been believed to show a wide variability in shell morphology. However, the population in Tanabe Bay, on the west coast of the Kii Peninsula, is known to consist of only two forms, which are distinguishable by the shape of nodules on the shell. As the first step toward a better understanding of shell morphology variation, the existence of reproductive isolation between these two forms was tested in Tanabe Bay using allozyme analysis. The two forms were confirmed to be isolated genetically; there are significant differences between them in allele frequencies at the 6 analysed loci.
Thais clavigera (Küster, 1858), a neogastropod snail, commonly inhabits intertidal rocky shores from southern Hokkaido, Japan, in the north to China in the south (Fujioka, 1986). In the Japanese main islands, this species is abundantly found on the Pacific coast and around the Seto Inland Sea. It seems absent in Okinawan tropical waters (Kubo and Kurozumi, 1995).
Thais clavigera has a robust shell with many nodules on the surface. The size and shape of the nodules vary among individuals. At Asamushi (Mutsu Bay), Shimoda and Dougashima (Izu Peninsula), several localities along the coasts of Ise and Mikawa Bays (Central Japan), and Ushimado (Seto Inland Sea), each local population consists of a majority of individuals with low or almost indiscernible nodules and a minority with well developed ones (Hayashi, in preparation). On the other hand, Abe (1985) found that the local population in Tanabe Bay on the west coast of the Kii Peninsula consisted only of nodulose individuals, of two forms designated C and P. Although occurring together, these forms are clearly distinguishable mainly by the different shape of their well developed nodules. The individuals of Form P are furnished with papillary nodules, while those of Form C have more sharply pointed, conical ones. Further, Abe (1985) found that the shell grew larger in Form C than in Form P. On the basis of these morphological differences, and especially his own field observation that cross-matings occurred only very rarely between the two forms, Abe (1985) claimed that the two forms should be recognized as distinct species. Later field observations and laboratory experiments revealed a clear difference in food preference between the two forms in larger individuals (Abe, 1989, 1994).
Abe's pioneering work was done exclusively in Tanabe Bay. It is necessary to take account of other local populations, such as those mentioned above, for a better understanding of variation, differentiation, and speciation in the T. clavigera complex. As the fundamental first step, however, I have tried to clarify the genetic relationship between Forms C and P in Tanabe Bay in terms of population genetics, using allozyme analysis.
DISTRIBUTION OF FORMS C AND P IN TANABE BAY
The distribution of Forms C and P of T. clavigera was surveyed in the intertidal zone along the coast of Tanabe Bay at low spring tide in May, 1996 (Fig. 1). Of many visited sites, only the following were inhabited abundantly by T. clavigera: Sakai, Tenjin, Edzura, and Bansho-zaki. On the other hand, the density was very low at Takinai and Hatakejima Is. I failed to find this species in the inner part of the bay, except at Takinai. At all these 6 sites, all the specimens could be assigned to either form very easily. Form C exceeded Form P in number at all the sites, especially remarkably at Takinai and Hatakejima Is. in the inner part of the bay (Fig. 1).
Fossils of T. clavigera were recovered in 1981 in the course of an archaeological excavation on the premises of the Seto Marine Biological Laboratory of Kyoto University, situated close to Bansho-zaki mentioned above, and deposited in the Center for Archaeological Operations of Kyoto University. Of the specimens from the layer estimated to have been deposited 4000 years ago (Niwa, 1977), 31 almost complete snails were easily classified as 20 individuals of Form C and 11 of Form P. The relative abundance of the forms was nearly the same as that of the extant population at Bansho-zaki (Fig. 1). Further attempts to get fossils of T. clavigera around Tanabe Bay were unsuccessful.
MATERIALS AND METHODS FOR ALLOZYME ANALYSIS
Specimens of Forms C and P of T. clavigera were collected from Bansho-zaki (the sites of Abe's earlier investigations) in June, 1994, and Sakai, Minabe-cho, on the opposite shore of Tanabe Bay (Fig. 1), in July, 1997. They were transported alive to the laboratory, and soon frozen and stored at −80°C until required for use. Adult specimens larger than 15 mm in shell length (Nakano and Nagoshi, 1980) were used, to avoid a possible change in allozyme banding pattern with growth (Markert and Ursprung, 1962).
Each specimen was thawed at room temperature. A small piece of filter paper was inserted into the cut foot muscle for 5 to 10 min to soak up fluids from the tissue. The pieces of filter paper from different individuals were inserted immediately into the origin of a horizontal 12% starch gel prepared according to Numachi's (1983, 1989) method. Enzymes were assayed using two different buffer media: Trisborate-EDTA buffer, pH 8.7, for superoxide dismutase; and citrateaminopropyl-diethanol-amine buffer, pH 7.0, for the other enzymes. Electrophoresis was conducted for 3 hours, with an ice bag to cool the gel field. Enzymes were resolved according to the staining methods of Numachi (1983, 1989). The enzymic systems studied were: isocitrate dehydrogenase (ICD, EC 188.8.131.52), malate dehydrogenase (MDH, EC 184.108.40.206), malic enzyme (ME, EC 220.127.116.11), 6-phosphogluconate dehydrogenase (6PGD, EC 18.104.22.168), phosphoglucomutase (PGM, EC 22.214.171.124), and superoxide dismutase (SOD, EC 126.96.36.199).
Samples of Thais bronni (Dunker, 1860) collected from Sugashima, Toba, Mie Prefecture, in July, 1995, were analysed as an outgroup using the same methods and enzymic systems employed for Forms C and P of T. clavigera.
Genetic differentiation of Forms C and P
The allele frequencies of 10 loci of 6 enzymes were analyzed for Forms C and P (Table 1). Only one locus, ME, was revealed to be monomorphic and the others were polymorphic.
Allele frequencies in the two forms of T. clavigera and in T. bronni. Numbers in parentheses are the sample sizes for each locus: Isocitrate dehydrogenase (ICD), malate dehydrogenase (MDH), malic enzyme (ME), 6-phosphogluconate dehydrogenase (6PGD), phosphoglucomutase (PGM), and superoxide dismutase (SOD)
The null hypothesis to test was the absence of differences in the frequency of each locus between the two forms at the two sites. The chi-square values were calculated following Nei (1990). Table 2 clearly shows that the null hypothesis was refuted, with probabilities of less than 5%, for the six loci ICD, MDH-1, MDH-2, PGM-1, SOD-1, and SOD-2. As for sympatric occurrence of the two forms, Forms C and P at Bansho-zaki were genetically separate at 3 loci, MDH-1, PGM-1, and SOD-2; the forms at Sakai were discriminated from each other at 3 loci, MDH-1, MDH-2, and SOD-2. On the other hand, for allopatric populations of the two forms, Form C at Bansho-zaki and Form P at Sakai were distinguishable at 4 loci, MDH-1, PGM-1, SOD-1, and SOD-2, while Form P at Bansho-zaki and Form C at Sakai were separate at 2 loci, ICD and SOD-2. Consequently, Forms C and P in Tanabe Bay were suggested to be isolated genetically from each other.
Pairwise chi-square values of polymorphic loci among the two forms of T. clavigera of two localities and T. bronni. Numbers in parenthese are the degrees of freedom for each locus. * and ** mean that the chi-square values for the loci have probabilities of less than 0.05 and 0.01, respectively. PGM-2 was not detected in T. bronni
Meanwhile, no genetic differentiation was detected between allopatric populations of the same form, i.e. between the Bansho-zaki and Sakai populations of Form C or Form P at any locus, including SOD-2. Especially at this locus, Forms C and Form P were clearly distinct from each other at each of these sites.
The chi-square values for 10 enzymes between T. clavigera and T. bronni indicate that the two species are clearly separated from each other; PGM-2 was undetected in T. bronni and thus was omitted from the analysis (Table 2).
Genetic distance between Forms C and P
In order to quantify the degree of genetic differentiation among Forms C and P and T. bronni, the genetic identity (I) and genetic distance (D) were calculated by the method of Nei (1990) from the allele frequency data of Table 1. Table 3 shows the matrices of I and D based on 9 loci (excepting PGM-2, which was not detected in T. bronni). The D values between Forms C and P (0.0132–0.0221) fall within the empirical value range among local populations of a species (Nei, 1990). The D values between the same forms (0.0035–0.0075) were slightly smaller than those found between forms. On the other hand, the D values between T. clavigera and T. bronni were 0.1161–0.1175, corresponding to the empirical value range among different species (Nei, 1990).
Genetic identities (above diagonal) and genetic distances (below diagonal) among Forms C and P of T. clavigera and T. bronni
Fig. 2 shows the dendrogram among Forms C and P from Bansho-zaki and Sakai and the outgroup T. bronni, reconstructed from Nei's genetic distance data by PHYLIP 3.57 computer software's UPGMA clustering method (Felsenstein, 1995). The Bansho-zaki and Sakai populations of each form are the closest relatives in T. clavigera.
This study has shown that the sympatric Forms C and P are genetically isolated in Tanabe Bay. This result is consistent with Abe's (1985) field observation that cross-mating between them was found very rarely at Bansho-zaki, in the bay, and it supports Abe's (1985) opinion that Forms C and P should be regarded as distinct species. It also supports Abe's (1994) suggestion that the difference in food preference in larger individuals between the two forms may have genetic basis. Abe's (1989, 1994) discovery of this difference is well understandable in terms of the diversification in foraging strategies among sibling species.
I tried to find morphological differences in the soft parts between Forms C and P of Tanabe Bay, but so far in vain. I examined the detailed structure of the radula and the shape and coloration of the tentacles, penis, stomach, and intestine. Imposex, i.e., the occurrence of male sex organs in females, was always detected in the present specimens. This phenomenon is known to be very common in the Japanese population of T. clavigera (Nakano and Nishiwaki, 1992; Horiguchi et al., 1994).
Abe (1985, 1994) discussed nomenclature and concluded that Form C should be assigned to Purpura clavigera Küster, and Form P to P. problematica Baker (1891). His reasoning is difficult to follow. I will discuss this problem elsewhere in detail, after the taxonomic status of local populations of T. clavigera outside Tanabe Bay is clarified fully in relation to the two forms in the bay. Tanabe Bay is not the type locality of either P. clavigera or P. problematica.
A high variation in shell morphology has been known in several species of Thais. Such variability may be caused genetically in some cases or environmentally in others. For example, the shell color variation of T. emarginata (Deshayes, 1839) is controlled by a single gene (Palmer, 1985). Also, the variation in the size and shape of the shell in T. lapillus (Linnaeus, 1758) is due to environmental effects, while its shell color and sculpture are attributed to genetic factors (Colton, 1922). As the genetically distinct Forms C and P of T. clavigera live sympatrically in Tanabe Bay, it can be naturally supposed that they are influenced by the same environmental factors. It follows that the shape of the nodules is genetically determined in Tanabe Bay.
Tanabe Bay is now inhabited only by Forms C and P and has probably been so for at least 4000 years, as shown by the examined fossils. On the other hand, my field survey has revealed that many local populations elsewhere comprise a majority (e.g., ca. 80%) of individuals with obscure nodules, and only a minority with distinct conical or papillary nodules similar to those of Forms C or P. On the basis of this consideration, the genetic relationships among Forms C and P in Tanabe Bay and the several morphs seen in other local populations should be clarified.
I express special thanks to Prof. T. Nishikawa of Nagoya University for kind advice throughout my research and for his critical reading of the manuscript. Thanks are also due to Dr. N. Abe of Osaka International College for Women for useful information and discussion and to Dr. M. J. Grygier of the Lake Biwa Museum for reading the manuscript critically. I am most grateful to Prof. A. Hino and Mr. H. Takizawa of Kanagawa University, Dr. R. Ueshima of the University of Tokyo, and Dr. T. Kobayashi and Dr. T. Okazaki of the National Institute of Aquaculture for valuable technical advice; to Prof. E. Harada and the staff of the Seto Marine Biological Laboratory, Kyoto University, Dr. Y. Shimizu of the Center for Archaeological Operations, Kyoto University, and Prof. H. Hayashi and the staff of the Sugashima Marine Biological Laboratory, Nagoya University, for their aid in collecting materials.