QUANTITATIVE BEACH RESEARCH ON WASHED-UP ASSEMBLAGES OF COWRY-SHELLS IN THE INDO-WEST PACIFIC REGION

Willem Krommenhoek Ph.D.

Usually collectors of shells washed-up on beaches just try to improve their collection by adding fresh specimens of as many as species as possible. When they are seriously engaged in collecting, they will look up the right name and label it accordingly. Many cupboards are filled with shells obtained this way during collecting trips. Of course, this is a fair thing to do and it provides a rewarding hobby, but there is so much more to do with beach found shells.

Since the 1970s I have been collecting washed-up shells in Indonesia, Sri Lanka and the Seychelles, and from the start I tried to use the countless numbers of washed-up shells for other purposes than collecting as well. I thought it must be possible to use this material to solve some biological questions which otherwise remain unanswered. It was clear from the start that solving problems in the line of distribution, variation, taxonomy, genetics etc. requires large numbers of shells, and for that reason I restricted myself to cowries. On tropical beaches washed-up cowries happen to be the most abundant shells, and fortunately most species are easy to recognize, even when somewhat eroded.

Over the years I presented many short papers to Of Sea and Shore Magazine and others, dealing with different subjects concerning cowries. Now, after three decades it might be of interest to present this updated survey of my quantitative beach research.

1. THE COWRY: ANIMAL, HABITAT, DISTRIBUTION AND RARITY

1.1 The adult cowry in short

An adult cowry-shell has fully developed teeth on both lips and a thickened base. Once this stage is reached, the shell increases only in bulk as a result of deposition of nacre and will not grow in length anymore. On reaching maturity, the adult color pattern of spots, bands, reticulations or solid color is developed and the characteristic slit-like aperture becomes evident. The cowry reaches maturity within months and its lifespan ranges from a couple of years for the smaller species to probably ten or more years for the larger ones.

Inside the shell is the visceral hump, containing most of the organ systems of the cowry. When the animal extends from the shell, its mantle, foot and tentacles become visible. Interiorly the mantle forms the siphon. On lateral extensions near the basis of the tentacles well developed eyes are found. The animal is attached to the shell by the large columellar muscle. Only these external visible parts of the cowry will be mentioned in some detail.

The mantle tissue is thin and has several functions. It secretes the calcium carbonate shell; it lays down a protein framework in which the crystals of calcium carbonate lie; it deposits the color pattern; it repairs the shell, and it protects the glossy surface of the shell which otherwise would become dull if left exposed to the surrounding sea water. Moreover, mantle tissue contains glands which secret acid in about 25 species. The siphon, which is a specialized part of the mantle, is concerned with respiration. Water is drawn into the shell through the siphon and passes over the gill where oxygen is exchanged for carbon dioxide.

The foot is not only muscular, but also contains glands which secrete mucus. The animal moves over the substratum as a result of rhythmic muscular waves from posterior to anterior through the sole of the foot, resulting in a speed of about 5-6 cm per minute. Under stress cowries can autonomize portions of the foot.

The radula is a flexible structure with transverse rows of chitinous teeth, used for browsing the substrate. When worn away these teeth are protruded and replaced by new rows which are continuously produced. Radular teeth have distinctive patterns and their shape, size and structure serve taxonomic purposes.

Cowries are nocturnal animals. They spend the daytime in their hiding places in trenches and under stones and coral blocks. During the night they can be found in all biotopes of reefs. The figure below shows the conchological characters of the cowry shell (After Burgess, 1985).

1.2 The habitat of cowries

About two-thirds of all cowry species are associated with coral reefs, restricted to tropical waters and subtropical waters between 30 degrees N and 30 degrees S of the equator, and to depths of less than about 25 m. This study is restricted to reef associated species of the Indo-West Pacific region only.

Coral reefs provide a wide variety of habitats, ranging from the open reef front exposed to heavy wave action to shallow and quiet waters of lagoons. Ecological conditions vary considerably between these extremes, especially in factors like temperature and quality of the sea water. Like corals, cowries are also zoned on reefs. Some species are found only on the outer slope, while others are restricted to the outer edge of the reef where the waves break. Some are conspicuous on the reef flat and others are only known from the lagoon. A few species are found in different habitats, e.g. C. annulus and C. moneta can be found on reef flats, in the lagoons, and in the sea grass beds.

Using data from the Marshall Islands (Johnson, 1974, in Burgess, 1985) and from the Indian Ocean (Taylor, 1971) we can summarize as follows.

1.3 The numerical abundance of cowries in washed-up shell assemblages

On beaches behind fringing reefs, cowries are usually the most abundant shells in washed-up assemblages. Yet the Cypraeidae is not the biggest group of mollusks. To illustrate this fact, the results of studies in French Polynesia (In Sorokin, 1995) are quite convincing. On the reef flat covered with sand with rear corals, 117 species of mollusks were counted. Of this total number the percentages for species of the dominating taxa were: Mitridae 15%; Cerithidae 11%; Conidae 10%; Muricidae and related taxa 8%; Cypraeidae 7%; Terebridae 6%; Strombiidae 4%. But in regards to the total number of specimens the Cypraeidae were most numerous with 52%
Five dominating species formed 75% of the total number of mollusks. They were: Cypraea obvelata (closely related to the common C. annulus); Cerithium piperitum; Strombus mutabilis; Terebra sp.; and Vermetus maximus.

Three species out of the 117 were responsible for 90% of the total molluscan biomass: Cypraea obvelata; Tridacna maxima; and Vermetus maximus. In the coastal lagoon, the bottom of which was covered with sand and a mixture of detritus and rear corals, just the one Cypraea obvelata comprised 60% of the total number of specimens. Its number exceeded 100 per square meter.

According to Slimming and Jarret (1977), about 40 species of cowries are present in the Seychelles in the western part of the Indian Ocean. This number increases eastwards to about 59 in Indonesia (Dharma, 1988) and 55 in western Australia (Wells and Bryce, 1988). If we take into account that some Indonesian species are only reported from the Moluca Islands in East Indonesia, we may say that from west to east the number of cowry-species increases from 40 to 55 over the Indo-West Pacific region. This means an increase of 37% over a distance of 5500 km. Like the authors mentioned above, I consider these species as monogeneric, representing true species of the genus Cypraea.

In order to find out the most accurate location of the evolution center of cowries, I calculated the number of species reported to occur in various areas of the Indo-Pacific region. I used Burgess' "Cowries of the World" as a reference, considering all cowries belonging to the single genus Cypraea. For the geographical distribution of a species I used his information, together with my own observations. The Indo-Pacific region was divided into eleven areas of about the same size and the number of species occurring in each area was counted. These areas were: 1. West Indian Ocean (East Africa, Madagascar, Seychelles); 2: Northwest Indian Ocean (Red Sea, Gulf of Aden, Persian Gulf); 3: Central North Indian Ocean (S. India, Sri Lanka, Maldives); 4: East Indian Ocean (Indonesia, Malaysia, W. Australia); 5: West Pacific (Philippines); 6: North Pacific (Japan): 7: West Central Pacific (Moluccas, New Guinea); 8: Central Pacific (Guam, Marshall Is., Mariana Is.); 9: South Pacific ( E. Australia, New Britain, Solomon Is., New Hybrides, New Calidonia, Loyalty Is., Fiji, Tonga); 10: East Pacific (Samoa, Tahiti, Cook Is., Marquesas); 11: Northeast Pacific (Hawaii).

Almost 90 species of cowries were found to occur in the South Pacific area and from there the number of species decreases in all directions. This makes the South Pacific area the most likely center of evolution of cowries. The decrease in the number of species towards the periphery of the Indo-Pacific region is dramatic in all directions: 34% in the West Indian Ocean; 40% in the Northwest Indian Ocean; 36% in the Central North Indian Ocean; 18% in the West Pacific; 31% in the North Pacific; and 56% in the East Pacific. The more distant from the evolution center, the higher the percentage of decrease.

1.5 Locations of quantitative beach research and species list

Over the years much time was spent discovering undisturbed beaches needed for reliable quantitative beach research. Actually, only very few beaches match the required state of being undisturbed. To start with, sufficient numbers of shells must be washed-up, and both the number species and relative abundance of specimens is determined by the quality of the reef. When a beach is near a populated area, most likely local people have disturbed the delicate balance of the reef in their search for edible organisms. In many cases the reef flat and algal ridge are seriously impoverished and the effect of using poison and dynamite for fishing only speeds up this process. According to the Indonesian Institute of Science only 7-8% of the Indonesian reefs is still in the original state. A similar situation is likely to occur in the Philippines.

Therefore, the only places matching the demands of being undisturbed and uninhabited are to be found in nature reserves only. And there poaching is a potential disturbing factor. For these reasons only a few locations were chosen for quantitative beach research from the tens of beaches visited along the shores of the Indian Ocean. A short survey of these locations is given below.

Indonesia. During the 1970s and 80s several visits were made to the uninhabited and undisturbed beaches of Ujong Kulon Nat. Park and Triangulasi Beach in Baluran Nat. Park, respectively on the extreme western and eastern tips of Java, as well as to Nyang-Nyang beach on the southern tip of Bali. Incidental visits were made to beaches in West Java and to the islands of Madura, Lombok, Moyo, Penida, Lembongan, and Ceningan, situated east of Bali. Also beaches in South Sulawesi and the islands Salayar and Tembalongan, south of Sulawesi, were visited. Finally a number of small islands in the Spermonde Archipelago, west of Sulawesi, were visited.

Seaweed cultivation on the reefs of Penida Island, east of Bali

Disturbed beach as a result of growing ofar in South Sulawasri, Indonesia

Washed-up cowries on an uninhabited beach on Salayar Island, Indonesia

Washed-up cowries on an uninhabited beach in West Java, Indonesia

Collecting site in West Java, Indonesia

Collecting site in East Java, Indonesia

Sri Lanka. After exploring much of the coast of southern Sri Lanka, two locations were chosen for research purposes. The first spot was located 8 km north of the city of Galle in southwestern Sri Lanka, the second spot was near the village of Ranna, between Hambantota and Tangalle in the extreme south.

Collecting site in South Sri Lanka

Seychelles. Most collecting was done on the beaches of the small island of La Digue, situated 4 degrees S. and about 1400 km off the African coast and northeast of the Malagasy republic. The island is only 6 km long and 3 km wide and is surrounded by a fringed reef. Most collecting was done in the southern part. Even on this small island an increased disturbance of the balance of the ecology of the reef was observed over the years, as was indicated by the gradual impoverishment of both quantity and species richness of washed-up shell assemblages. Besides an increase of the local population, a major factor for this impoverishment is a booming shell industry, stripping the reef of all shells that could have some commercial value.

Collecting site in Seychelles

India. In the late 1970s the southeast coast of India was explored between Kodikkarai, north of Palk Bay, and Rameshwaram on Adam's Bridge, the isthmus between India and Sri Lanka. Only data from the Adam's Bridge area were used in this study.

Adam's bridge on the Indian side, South India

In all cases collecting was done during the months of July and August when strong monsoonal winds produce heavy wave action resulting in fair amounts of washed-up shell assemblages.

The list below shows the cowry-species found at the main collecting sites during the period 1977-1996. + has been found; - does not occur in the region; x occurs in the region, but has not been found. * includes C. leviathan which is not accepted by me; ** includes C. grayana which is not accepted by me; *** includes C. vredenburgi which in my opinion is only a form of C. pallida.

N.B. 1: In my opinion the occurrence of C. histrio and C. miliaris in West Java is very doubtful.
2: In my opinion East Java is not in the range of C. mappa.
3: The occurrence of C. lutea in West Java is very rare.
________________________________________________________

1.6 Distribution of cowry-shells on sandy beaches

Anyone familiar with collecting shells on sandy beaches knows that washed-up shells and debris are distributed in irregular patterns resulting from the interaction of local currents and the steepness of the beach. The distribution also changes with time: what is a good collecting site one day may be an empty spot the next. Moreover, the season determines what will be found on the beach: strong monsoonal winds result in heavy wave action, which in turn results in deposition of different material. Generally one may say that sediment, shells and debris are transported by interacting currents, waves and winds, resulting in transport parallel, perpendicular or diagonal to the coast. A short analysis of the different elements that play a role in the distribution of washed-up shell assemblages on tropical sandy beaches is given below.

Beach. Most tropical beaches behind a reef are characterized by an abundance of organic carbonate remains, ranging from tiny eroded particles of coral to large parts of debris. The particles of such coral sand beaches are usually well rounded due to the softness of the material. Other particles are composed of mineral grains derived from various rocks. Quartz usually dominates because of its hardness. Magnified, these grains are angular in shape, the quartz being light in color and other minerals reddish or black, depending on the type of rock from which they originated.

Locally a carbonate beach can turn into a radiolarian beach, consisting of billions of round grains, about 1-2 mm in diameter and composed of pure silica. They are the remains of an astronomical number of protozoans which store silica in their body walls. Small openings in the surface, through which hairs of protoplasm protruded, are still visible under magnification. Radiolarian beaches are easily recognized. Where a carbonate beach is usually solid, radiolarian beaches are very loose, making a person sinking foot deep in it as a result of the loose packing of the particles.

The upper section of many tropical beaches consists of a horizontal or landwards sloping surface called the berm. This is a zone of vertical accretion formed by backwash deposition, its height being limited by the upper limit of swash. On the seaward side the berm crest gives way to the beach face with a slope of about 2 degrees in fine sand. Although the slope is primarily determined by the coarseness of the sediments, wave height and steepness also play a role in this. Steeper and higher waves are associated with gentler gradients as these waves generate more backwash.

Where the beach gradient is shallow, a submerged long shore bar can be observed, running parallel with the beach and separated from it by a trough. Sometimes several of these bars can be present. They seem to develop in response to the action of braking waves.
Finally, rhythmic features along the beach may be observed, the smaller being beach cusps, crescent shaped indentations lying parallel with the shore on the upper beach face. These cusps may be initiated by irregularities in form along the beach which are subsequently enlarged by swash, with the eroded material being deposited on either side of the irregularity by the backwash until equilibrium is attained and the sand particles move in a close circle.

Waves. In deep water, waves are called oscillatory since the individual water particles move in a circular motion. This motion declines rapidly downwards and at a depth of about one half of the wave length, known as the wave base, there is hardly any movement left. As waves move towards shallower water and the water depth decreases to that of the wave base, the sea floor starts to interfere with the circular motion and the movement of water particles becomes more elliptical. Now forward movement of the water particles becomes important and the original oscillatory waves become translatory waves. With further decreasing water depth, the wave height increases and the wave becomes steeper. Eventually the wave is over steepened to the stage where it breaks as its crest crashes forwards, producing plunging breakers and creating surf. Once the wave form has been destroyed, the remaining water moves up to the shore as swash and returns under the force of gravity as backwash.

Incoming waves always tend to be roughly parallel with the shore line. This is the result of wave refraction. If we imagine a wave approaching a straight coast obliquely, those parts of the wave which reach shallow water first, will slow down because decrease in water depth to less than half the depth off wave base, reduces speed. This part of the wave will be caught by the other parts still in deeper water. The result is that incoming waves always tend to be parallel with the shore line as we all know by experience.

Currents and sedimentation. Next, we have to understand how waves can generate currents near the coast line. It is obvious that these waves are responsible for the actual transport and eventual sedimentation of debris and shells. It is well known to every beach observer that the backwash is normally confined to more or less evenly spaced zones forming rip currents which are fed by water moving parallel to the beach in long shore currents. The resulting circular or cellular circulation pattern is caused by variations in wave height of waves parallel to the shore. Water from higher waves will produce more swash and returns at the points where the lower waves have broken. It is believed that these variations in wave height and the subsequent cellular circulation, are the result of resonance between waves arriving at the shore and those reflected. This resonance might be responsible for raising the height of breakers at more or less regular intervals. Whatever the exact explanation, once established, rip currents and long shore currents are to some extend self-sustaining as a result of small modifications in the form of the beach. This explains why shells are found washed-up in irregular patterns.

But there is still more. Where waves strike the shore obliquely, the long shore currents will be in predominantly one direction and as a result of that, sediment and shells are transported along the shore. The actual effect will depend on the wave energy and the angle at which waves strike the coast. This explains why sedimentation normally changes with the seasons. When the prevailing monsoonal winds change their direction, they also change wave height and wave energy. The angle at which the waves arrive will also change. The result of this is a different sedimentation pattern.

Finally, where waves steepness and therefore wave energy is low, there is the effect of beach drift. Sediments and shells move along the beach for a short distance in each cycle of swash and backwash. This occurs when sediment moves obliquely up to the beach and is returned down the beach as a result of gravity.

From the above it will be obvious that shells cannot be found washed-up in a regular way.

1.7 The concept of rarity

Among collectors of cowries it is popular to express the rarity of a species on a scale from 1 to 20 as presented by Burgess (1985). However, that scale lacks any foundation or definition, and for several reasons mentioned below it is better to abandon it. In my experience, a species has no fixed rarity at all. As will be shown in the next chapter, several species have a clustered distribution rather than occurring homogeneously over their respective ranges. Being frequent at one location, the same species may be found only occasionally or rarely at another, not necessarily far away, as I found in Indonesia.

Secondly, some species gradually change in frequency from east to west or vice versa over the Indian Ocean, as I discovered after statistical analysis of large samples from the Seychelles and Indonesia. Both regions are located 4-6 degrees S, separated by 5500 km of ocean.

Thirdly, the frequency and hence rarity of certain species vary with the season, giving a different result when data for different seasons are compared.

It is well known that the washed-up assemblage of dead material, the thanatocoenossis, differs from the composition of the living biocoenosis, being the result of transport and deposition. Therefore, the relative frequency of beach found species cannot be expected to match the real frequency and rarity of it. However, as it is not only difficult to establish the frequency of a species in the living biocoenosis because of the nocturnal habit of cowries, but virtually impossible to follow fluctuations through space and time, we can only use the thanatocoenosis as a source of information.

Lorenz and Hubert (1993) noted that all past attempts to quantify abundance/scarcity had failed, and they introduced an eight-step scale of very common, common, moderately common, uncommon, moderately rare, rare, very rare and extremely rare. However, the terms uncommon and moderately rare are for all practical purposes synonymous, the distinction between common and moderately common is too vague to be meaningful, and very common means abundant. Therefore, I recommend using six categories only: abundant, frequent, common, occasional, rare and very rare. These terms, of course, also have room for doubt, but some scale has to be used, and keeping it as simple as possible can reduce the confusion about rarity.

2. GENERAL SUBJECTS

2.1 Recover percentage of total number of species

At all main locations where collecting was done, over 90% of the total number of species occurring in that region was found washed-up on beaches. Surprisingly it only took a few days to reach this percentage, provided the washed-up assemblage was large enough. The highest score of 94% was found at Grajagan Bay in Baluran Nat. Park in East Java on an undisturbed beach.

The relation between the size of the sample, i.e. the total number of beach-found identifiable shells, and the percentage of recovery of the species occurring in a region is an asymptotic approach to a theoretical 100%. In samples of 2000 shells, about 80% of the number of species occurring in the region was included, increasing to 90% when the sample increases to 6000-7000, and reaching 94% in a sample of 20.000.

It is obvious that those species living in deeper places beyond the reef are unlikely to be washed-up.

2.2 Percentage of identifiable shells in a washed-up assemblage

After collecting tens of thousands of washed-up cowries, it was found that generally speaking about 1% of the specimens are fresh and undamaged. Shortly after death erosion of the shell begins, resulting in a dull surface, pigment reduction and finally lack of color. This is the result of abrasion by mineral grains and other moving objects, due to which the shell gradually loses its gloss and shine.

At first it is not easy to identify an eroded shell, but the more specimens one has seen, the better one understands how the gradual change in appearance takes place. After some time of practice, it was possible to identify up to 90% of the beach-found cowries, which is a surprisingly high percentage. The remaining ones are abraded beyond recognition. However, this is an overall estimation as some species lose their characteristics more rapidly than others.

2.3 The rate of erosion of washed-up cowry-shells

Ever since I observed how washed-up cowries lose their gloss and shine, I wondered how long it takes to transform an undamaged shining cowry into a dull shell. To find an answer to this question, I collected 600 specimens of Cypraea ocellata in south Sri Lanka in all stages of wear and erosion. The species is very common in Sri Lanka and large numbers of shells are easily available. I observed at that location how during incoming tide a strong backwash for about 4-5 seconds vigorously drags seaward all the debris and shells the swash had brought ashore. It is during this backwash that shells undergo countless collisions and erosion takes place. This process is repeated 5-6 times every minute for a period of about two hours. During other phases of the tide there is far less action and when the tide is in or out, the particles come to rest. From this observation I came to an estimated 45 minutes every day (there is one tidal cycle per day) during which the eroding action takes place.

Next I classified the 600 specimens according to the following scale: intact and glossy; intact and dull; outer layer of the shell with the characteristic spots damaged and worn; outer layer completely eroded. The ratio for these categories worked out to be 1 : 4 : 16 : 19. This means that for each intact glossy shell there are four dull ones, 16 with the outer layer damaged and 19 or possibly more with the outer layer of the shell completely worn down in beach-found shell assemblages of C. ocellata.

After that I selected intact and glossy specimens of C. ocellata, C. interrupta and C. moneta and put them in a plastic container half full with wet eroded specimens and debris. To imitate the effect of drag caused by the backwash, I started turning, shaking and rotating the container for some time. After 3 minutes of mild shaking, turning and rotating, the surface of the nacre showed many fine scratches in all three species when examined with a magnifying glass. After 5 minutes more microscopic scratches were observed, the shell was still shining. After 10 minutes the whole surface of the shell was covered with fine scratches and small punctures, the shell had still some shine. After 15 minutes the shells of C. ocellata became dull, the other species had still some shine. After 20 minutes all shells were dull. After 30 minutes the complete shiny outer layer of C. ocellata specimens had gone. The layer with the oscillated spots started to be eroded.

After 45 minutes, the sulcus in C. ocellata was widening as a result of the loss of material; in C. interrupta the layer with the fine brown spots was eroded for about 50%.

After one hour of imitating the effect of the backwash it was observed that on the dorsum of C. ocellata shells the sulcus widened further to a mm or more, gradually exposing the deeper green-blue layers of the shell. C. interrupta was now rapidly losing the layer with the fine brown spots, exposing deeper bluish layers with three transverse brown bands. C. moneta was now losing its fine yellow color and most of the orange ring which is so common in this species in Sri Lanka.

From the above mentioned data we may conclude that after only a few days cowry shells have lost their outer layers and are transformed into worn shells without any characteristics except their shape, basal ring and teeth. As the inner layers of the shell usually are rather thick, it will possibly take weeks to erode the shell until cracks and openings appear and eventually the dorsal part has gone, leaving only the base with the teeth. However, the length of this final stage of erosion will depend on the species as well on the age of the shell, as these factors determine the final thickness of it.

2.4 Beach-found shells as evidence for the presence of a species

From my observations I reached the conclusion that freshly washed-up undamaged shells cannot have been transported over a considerable distance after the death of the animal. As has been shown in the previous paragraph, fresh shells washed-up on beaches lose their gloss and shine within 24 hours as a result of countless collisions with other shells and debris, mainly during backwash when the tide comes in. Therefore, a fresh shell with little or no abrasion cannot be expected to have arrived from a considerable distance, and so its presence can be regarded as evidence for the presence of the species in the region.

2.5 Notes on washed-up cowry-shells and zoögeography

In this section information about rarity and distribution as it is available in the literature is commented, based on personal field observations. For general information on distribution of cowries the work of Schilder (1965) and Burgess (1985) was taken as reference. The splitting taxonomy practiced by Lorenz and Hubert (1993), in which a valid species is divided in several new taxons, is not taken into account. For local information about distribution the data given by Slimming and Jarret (1970) for the Seychelles; Dharma (1988) for Indonesia; and Wells and Bryce (1988) for West Australia were used.

Although Schilder's publication is 20 years older than Burgess', it contains more detailed information and in many cases larger distribution areas. Unfortunately, Schilder did not mention any sources of information. He introduced a system of classification of faunas based on a code of letters and numbers resulting in 160 areas. Burgess ignored this system and introduced maps which are much easier to read. However, although much literature is cited and personal comments of collectors are included, in many cases the ranges of species are smaller than in Schilders' work, without giving any reason. But also in other publications the information about the distribution of cowry species is inaccurate and sometimes contradictory. Therefore, systematic collecting washed-up cowries can provide reliable new information about the accurate zoögeography of this group. As will be shown below, collecting washed-up shells resulted 13 times in the proven evidence of a species in an area where it was not known to occur. For a number of species it was also found that where it is very common at one place, it may be almost absent in another location with similar conditions of latitude, type of beach, steepness of the beach, wave action etc. This suggests that a number of species is characterized by a clustered distribution rather than being distributed homogeneously over its range. On the other hand, it is also possible that irregular distribution results from unknown patterns of transport and deposition. Therefore, general statements on distribution and rarity are rather meaningless.

Field observations for most of the species occurring in the Indo-West Pacific region are given below.

C. annulus L. 1758 is usually number one in frequency in both the Seychelles and Indonesia and in my opinion the species is even more abundant than C. moneta. In south Sri Lanka it was only occasionally found. Burgess' map must be extended to the north to include Sri Lanka, like Schilder did. The species varies mainly in the color of the dorsum, ranging from creamy white to greenish gray and, especially in younger specimens, with a bluish or occasional light brown shine. Very light forms were observed in the Seychelles only.

C. arabica L. 1758 was found frequently in Indonesia and occasionally in the Seychelles and south Sri Lanka. According to Wells and Bryce it is uncommon in Western Australia. This pattern suggests a gradient from Indonesia to the limits of its range in the west and south. Species from the Seychelles are characterized by heavy callus on the margin. I also noticed specimens resembling C. depressa but with reticulations of the C. arabica type. According to the information given by Schilder and Wells and Bryce, Burgess' distribution map must be extended to the north as well as south.

C. argus L. 1758 was found occasionally in the Seychelles in a pigment-reduced form with yellow rings and bands. In Indonesia and on Sri Lanka it had its normal pigmentation and color, with its frequency varying from locally frequent to rare, supporting my idea of clustered distribution. For Western Australia it is reported to be uncommon. Consequently, rarity varies from frequent to rare.

C. asellus L. 1758 was found in fair numbers in both the Seychelles and Indonesia, being frequent in the last and occasional in the first region. It was not found in Sri Lanka, because in my opinion that is out of the range of the species. According to Wells and Bryce it is common in Western Australia. This is another example of a gradual decrease in frequency towards the limits of its geographic range. The species is constant in shape and pattern of the dorsal bands, varying from brown to orangey red, the latter mainly occurring in the Seychelles. Burgess' distribution map needs to be extended to the south to cover Western Australia and to be withdrawn from Sri Lanka.

C. boiviniii Kiener 1843 is one of the few examples of cobalt blue occurring in nature. Although there are many specimens with white and brown spots only, some have beautiful cobalt blue spots, each surrounded by a brown ring and in some cases with blue between the spots as well. The species is variable in the amount of blue rather than in shape. It was found in fair numbers in East Java in a multitude of color shades, less frequently in Bali and Lombok and not at all in Sulawesi. This observation suggests a gradation in frequency over a small area. The species has a limited distribution in the eastern part of the Indian Ocean and was not observed in Sri Lanka. It is not recorded in Western Australia either.

C. caputserpentis L. 1758 was found frequently in the Seychelles, south Sri Lanka and Indonesia. At all location specimens differ a great deal in the amount of callus. Young individuals have an embryonal band and lack the characteristic white spots that cause a reticulate pattern on the dorsum. Remarkable in this respect is that occasionally young specimens are found together with specimens showing all characteristics of adult ones but smaller in size. The color usually varies from chocolate to dark brown, but occasionally much lighter.

C. carneola L. 1758 was found in the Seychelles, Sri Lanka and Indonesia, being frequent to common at all places. The color of the shell varies from deep carneolan bands on a pinkish gray dorsum to a yellow dorsum with slightly darker yellow bands. Light colored specimens were found mainly on the Seychelles. The shape varies from cylindrical to elongated ovoid. Because of these gradual variations, I doubt the value of splitting the carneola group into four or five species, as Burgess has done.

C. caurica L. 1758 was found abundantly on the Seychelles, frequently in Indonesia and incidentally in south Sri Lanka, indicating a frequency decrease from west to east. Pigment gradation is very prominent in this species, with shells ranging from normal to very light and almost albino in the Seychelles, where they also have more callus.

C. cernica Sowerby 1870 is described as rare and occurring in deep water, so washing-up on beaches is very unlikely. Moreover, Dharma mentions it only for the Moluccas in East Indonesia. This makes Burgess' distribution map unlikely and explains why I never found this species on beaches.

C. chinensis Gray 1825 was found occasionally in the Seychelles, Sri Lanka and Indonesia, making it one of those species that seem to be distributed homogeneously over its range. Burgess' distribution map must be extended northwards to include Sri Lanka, like Schilder did.

C. cicercula L. 1758 is one of two closely related Indo-West Pacific species, found in the Seychelles, Sri Lanka and Indonesia. The species is supposed to have a characteristic spire blotch, but a fair number of the specimens collected were creamy white and showed only a small depression. The dorsal sulcus, which is supposed to characterize C. bistronotata Schilder & Schilder 1937, was always present. The dorsum was always granulate, although that is not characteristic according to Burgess. Nevertheless, in my opinion the species is fairly constant and can easily be identified by the spire blotch and/or a depression in the same place. In Indonesia the species is frequent to common, decreasing to occasional and rare at the limits of its range like in the Seychelles, Sri Lanka and Western Australia. Burgess' distribution map must be extended northwards to include Sri Lanka.

C. clandestina L. 1758 was found frequently in the Seychelles, Sri Lanka and Indonesia. There is considerable variation in coloration, ranging from three light to dark coffee-colored bands crossing the dorsum to a bluish gray dorsum. In all cases the fine zigzag lines over the dorsum make it easy to identify. The shape is rather constant. Burgess mentions the species as occurring in Western Australia, but Wells and Bryce do not.

C. contaminata Sowerby 1832 is supposed to occur in the Seychelles, Sri Lanka, Indonesia and Western Australia according to Burgess. However, I never found a single specimen. This is not surprising because in the Seychelles the species has never been observed and according to Schilder the species does not occur in southern Indonesia. Therefore, I doubt very much the distribution given by Burgess.

C. cribraria L. 1758 was found in the Seychelles, Sri Lanka and Indonesia, being common to occasional at all locations. The shells are always thin and delicate, with a coloration that is easily eroded, leaving a white shell. Shape and coloration are remarkably constant, the only variation being the amount of callus.

C. depressa Gray 1824 is a species with large fluctuations in its frequency, being occasional in the Seychelles, common in Indonesia and occasional again in Western Australia. In Indonesia the frequency differs considerably at the various locations visited, suggesting a clustered distribution. Burgess' distribution map must be extended over south Indonesia to Western Australia as finds on beaches on the Lesser Sunda islands and data from Schilder and Wells and Bryce indicate.

C. diliculum Reeve 1845 occurs only in the western part of the Indian Ocean and was incidentally found in the Seychelles where it is described as rare.

C. eglantina Duclos 1833 occurs only in the eastern part of the Indian Ocean and was found occasionally in Indonesia. It is said to be common in Western Australia.

C. erosa L. 1758 was found to be common in the Seychelles and Indonesia and occasional in Sri Lanka. Specimens from the Seychelles are large with a fair amount of callus and are predominantly yellow to pale brown on the dorsum, with cream spots and a smaller number of chocolate spots with white centers. The Indonesian specimens have less callus and are markedly darker due to the addition of bluish-green shades on the dorsum,

C, errones L. 1758 only occurs in the eastern part of the Indian Ocean, being common in Indonesia and, according to Wells and Bryce, in Western Australia. The bluish-green dorsum has fine brown dotting of varying density and a dorsal blotch of varying size. Washed-up shells easily erode, exposing the bluish-green basic color in three darker and two lighter transverse bands. Brown anterior spots are often present, but specimens with only one spot or no spots at all also occur.

C. felina Gmelin 1791 shows a very prominent gradation in frequency over its range, In Java, Bali and south Sri Lanka it is common, but on Lombok which is east of Bali, it is found occasionally, and Wells and Bryce describe it as rare in Western Australia. On the Seychelles the species is occasional to rare. This is another good example of east-west and north-south gradation in frequency. Burgess' distribution map must be extended southwards to cover Western Australia.

C. fimbriata Gmelin 1791 was found in the Seychelles, Sri Lanka and Indonesia. At all places it was common and sometimes abundant, like in south Bali where it was used for making necklaces. Specimens vary in size rather than in shape or coloration. In Western Australia it is also said to be common. Burgess' map must be extended northwards over Sri Lanka, like Schilder did.

C. gangranosa Dillwyn 1817. If we adhere to Burgess' criteria for distinguishing between C. gangranosa and the related C. labrolineata Gaskoin 1849, then I found quite a few specimens of the former on Java. C. labrolineata should be recognized by its dorsum with discrete white circular spots and no dark spots, and by the absence of color in the canals. However, I have collected quite a few specimens with colored canals and a variety of brown spots on the dorsum amidst discrete white spots. Some of the brown spots have a bluish center resembling C. boivinii, but the blue is never so bright. I did not find C. gangranosa on the Seychelles, and neither do Slimming and Jarret mention it from there. It was found in south Sri Lanka where I regard it as occasional. Burgess' distribution map must be revised considerably since the species is absent in the Seychelles and present on Sri Lanka, as noted already by Schilder.

C. globulus L. 1758 is the most common species of the globular cowries. It was found occasionally in the Seychelles and more frequently in Indonesia. It can be easily identified by the smooth dorsum and the absence of a dorsal sulcus. The fine brown spots are characteristic. As in the related C. cicercula , light colored and even creamy white specimens occur in both the Seychelles and Indonesia, where I consider the species as common. In Sri Lanka it is occasional and Wells and Bryce describe it as such in Western Australia. The species is a good example of decreasing frequency to the north and south from the distribution center. The related C. bistrinotata differs in detail, and was separated from the other globular cowries by the Schilders in 1937. The species is said to resemble C. cicercula very closely, but with characteristic basilar spots. However, I never found such a specimen. What I did find were specimens of C. globulus with four basilar spots, but that is not mentioned as a diagnostic character.

Since both the Seychelles and Indonesia are outside the range of this doubtful C. cicercula, I cannot give a definite opinion.

C. gracilis Gaskoin 1849 was found in fair numbers in Sri Lanka and Indonesia where I consider it as being common. I did not find it on the Seychelles and neither Slimming and Jarret nor Schilder mention it from there. Burgess' distribution map is obviously incorrect here. In Indonesia dark specimens with a dark bluish-gray dorsum, a fair blotch of chocolate brown and numerous small brown spots were found.

C. grayana Schilder 1930 is a species I do not accept because, in my opinion, there are no reliable criteria for separating it from C. histrio Gmelin 1791. Close examination of many specimens indicated that none of the characteristics of C. grayana as mentioned by Burgess are constant and therefore the two species cannot be separated on conchological grounds.

C. helvola L. 1758 was found in large numbers in the Seychelles and Indonesia and incidentally in Sri Lanka. There is little variation in shape, but a lot of variation in color among specimens from the Seychelles and Indonesia. Apart from specimens with very little pigmentation, I noticed considerable variation in the size and number of brown spots as well as in the brightness of color. The light forms from the Seychelles show the characteristic spotting in shades of yellow only. I consider the species as frequent to abundant in the Seychelles, common in Indonesia and occasional in Sri Lanka. It is said to be common in Western Australia, indicating a decrease in frequency from west to east.

C. hirundo L. 1758 was found in large numbers in the Seychelles, Sri Lanka and Indonesia. It resembles C. kieneri Hidalgo 1906, but even worn specimens can be easily identified by the teeth. At one locality in the Seychelles, La Digue island, I collected a number of very light-colored specimens. As mentioned before, loss of pigment is found in other species on the Seychelles as well. I consider it occasional in the Seychelles and common to frequent in Sri Lanka and Indonesia.

C. histrio Gmelin 1791 was found frequently in the Seychelles. In Indonesia and Western Australia the species is said to be rare. This is an excellent example of decrease in frequency from west to east. In the Seychelles about 30% of the washed-up found specimens was so poorly pigmented that the shell only showed a light yellow pattern of reticulation in stead of the normal brown pattern.

C. interrupta Gray 1824 has a limited distribution and was found frequently in south Sri Lanka and occasionally in East Java and Lombok in Indonesia. It is not mentioned to occur in eastern Indonesia and Western Australia. It was not found in the Seychelles as this is outside the range of the species.

C. isabella L. 1758 was found frequently in the Seychelles, Sri Lanka and Indonesia. Although Burgess states that the species is constant in shape and color, I observed a lot of variation in color. Specimens from the Seychelles vary from very light creamy yellow with bright yellow terminal spots, through yellow with orange spots and bluish-gray with reddish-orange spots. The lighter forms dominate in the Seychelles. In Indonesia the reverse situation occurs, the bluish-gray type with reddish-orange terminal spots dominating the lighter forms. Specimens from either place vary in the number of longitudinal streaks. Two lighter bands running transversely over the dorsum were present in all forms except the very light ones from the Seychelles. Burgess' distribution map must be extended northwards to include Sri Lanka, as was also done by Schilder.

C. kieneri Hidalgo 1906 was found frequently in Sri Lanka and Indonesia and may be considered as common in Indonesia. Locally it can be very abundant. The species differs from the related C. hirundo L. 1758 by its teeth and more cylindrical shape. Like Burgess, I do not believe in local races of the species at each side of the Indian Ocean, because minor variations in shape and dorsal pattern occur at each side and within each population. Burgess' distribution map must be extended southward to include Western Australia where the species is said to be common as well.

C. labrolineata Gaskoin 1849 was found in central and south Indonesia only as its range does not extend over the Indian Ocean. As mentioned under C. gangranosa, distinction between these two species is not always possible because characters overlap, a situation noted also for C. globulus - C. bistrinotata and C. histrio - C. grayana.

C. lamarckii Gray 1825. Although Burgess lists the species as occurring in the Seychelles and Sri Lanka, I never found a single specimen. Slimming and Jarret mention it as rare in the Seychelles and unlikely to be found. Dharma mentions the species to occur in northern Indonesia only, which explains why I did not find it.

C. leviathan Schilder & Schilder 1937 is another species I do not accept. In my opinion it is a form of C. carneola L. 1758 with marginal nodules. In fact, this form was found in the Seychelles, on the East African coast and in Indonesia among normal C. carneola specimens. Even if the radula is different, as stated by Burgess, I do not feel specific distinction is warranted. Conchologically, leviathan-like specimens are not restricted to the area around Hawaii as stated by the Schilders, and therefore this type is certainly not a Hawaiian endemic species.

C. limacina Lamarck 1810 was found in fair numbers in the Seychelles and Indonesia, where I consider it respectively as occasional and common. There is less variation in color than in the related C. staphylea L. 1758, although light specimens were found both in the Seychelles and Indonesia. One specimen from Salayar island, a small island south of Sulawesi in central Indonesia, is creamy white without orange tips and looks like an albino.

C. lutea Gmelin 1791. Although Dharma lists this species as frequent in Indonesia, I found it only once in West Java. For Western Australia Wells and Bryce describe it as uncommon. Burgess' distribution map must be extended considerably to the west, which is also suggested by Schilder.

C. lynx L. 1758 was found frequently in the Seychelles and Indonesia, and occasionally in Sri Lanka. Especially in Indonesia I observed what I called a clustered distribution. Specimens do not differ in shape, but in the number and position of dark spots on the dorsum. Specimens with very little pigment were found in the Seychelles.

C. mappa L. 1758. Because of the relative rarity of the species, it was not found washed-up on beaches in the Seychelles and Sri Lanka. However, nice specimens were found occasionally in Java and Sulawesi, Burgess' map must be extended further south over Indonesia, which agrees with Schilder.

C. marginalis Dillwyn 1827. If a true species at all, it is restricted to the western side of the Indian Ocean. According to Burgess, it can be distinguished from the related C. poraria L. 1758 by the oscillated spots. Although Burgess mentions the species to occur in the Seychelles, I never found a single specimen. This is not surprising as Slimming & Jarret mention one oral report only of its occurrence there.

C. mariae Schilder 1927 is restricted to the eastern part of the Indo-West Pacific region. It is mentioned in this paper because of Burgess' statement that the species occurs in South Indonesia. However, most likely Burgess is not correct in this. I have never found a single specimen, and Schilder does not mention its occurrence there either. Dharma describes it as rare.

C. mauritiana L. 1758 was found in the Seychelles, Sri Lanka and Indonesia. This species clearly demonstrates its occurrence in clusters, being rather common in one place and rare or even absent in another. The species is most frequent in places with rough water.

C. microdon Gray 1926 should occur in the Seychelles, but I never found it. This is not surprising, as Slimming & Jarret admit that they have never seen a specimen from there either and that the information is from 'reliable sources' only. The species does not occur in South Indonesia, according to both Burgess and Schilder.

C. minoridens Melvill 1901 is one of the smallest cowries. I found it only in Java and Bali and not in the Seychelles. Burgess' map is incorrect here, as the species does not occur in the Seychelles according to both Slimming & Jarret and Schilder. For Indonesia I consider it as rare. Fresh beach-found specimens can be identified by the discrete brownish spots on the dorsum and bright red or orchid terminal spots. In my opinion the species is more ovoid than the related C. fimbriata.

C. miliaris Gmelin 1791. Although the species is described as common in Indonesia by Dharma, I found only one specimen on the small island of Puteran, east of Madura, which is northeast of Java. Burgess and Schilder do not agree on its distribution, but in my opinion the species is absent in South Indonesia. It does not occur in the Indian Ocean.

C. moneta L. 1758 competes with C. annulus for being the most abundant species. I found it in several variations. In the Seychelles the knobby form was found regularly, together with a golden-yellow form. Also specimens with an annulus-like ring were found. In Indonesia there was less variation in shape and color. In Sri Lanka the species was found less frequent, probably the result of being at the limits of its range. Burgess’ distribution map must be extended northwards to the Asian mainland and according to Schilder, southwards over Madagascar.

C. nucleus L. 1758 was found in fair numbers in the Seychelles, Sri Lanka and Indonesia. All specimens are remarkably constant in shape and color. However, the shell rapidly loses its brown gloss and turns into a dull bluish gray. Young specimens are smooth and only later the granulated structure and dorsal sulcus develop. In Indonesia the species is common, in the Seychelles occasional as it is in Western Australia according to Wells and Bryce.

C. ocellata L. 1758 has a limited distribution and was found on Sri Lanka and in South India only. Here it is frequent and at some places abundant. Whereas at many places in the Indo-West Pacific region C. annulus is the most frequent species, this position is taken by C. ocellata in south Sri Lanka and south India.

C. onyx L. 1758 is highly variable and consequently divided into several forms and variations. In the Seychelles the form adusta Lamarck 1810 was found occasionally. In Indonesia the species was found on Sulawesi only. Although Dharma describes it as frequent, in my opinion it is occasional to rare.

C. ovum Gmelin 1791. This species closely resembles C. errones L. 1758, from which it can be distinguished only by the yellow teeth. I fail to see Burgess' other character, a pyriform shape versus a more cylindrical shape in C. errones. If there is a difference in shape at all, then this difference is variable and not species-diagnostic. I found C. ovum incidentally in Indonesia where Dharma describes it as common. I presume the coloration of the teeth to fade away quickly after the death of the animal, making identification doubtful. However, it is my impression that in C. ovum the shell is thicker than in C. errones. Surprisingly, Wells and Bryce do not mention the species in Western Australia.

C. pallida Gray 1824. Several freshly washed-up specimens were collected in East Java and south Bali. Although Dharma considers the species as being rare, I found it common in East Java and occasional in south Bali. It was found incidentally in Sri Lanka, but not on the Seychelles, as this is out of its range. The related C. vredenburgi Schilder 1929 described to occur in Java only, was either not found or was not recognized as a different species.

C. poraria L. 1758 was found occasionally in the Seychelles, Sri Lanka and Indonesia, with a somewhat higher frequency in the Seychelles. Burgess' distribution map must be extended further south over Western Australia where it is known to occur occasionally according to Wells and Bryce.

C. punctata L. 1758 was occasionally found in the Seychelles, Sri Lanka and Indonesia. The characteristic brown spots are easily eroded, leaving a dull white shell, a situation like in C. cribraria.

C. pyriformis Gray 1824 was surprisingly not found in southern Indonesia, although Dharma describes it as being frequent in Indonesia. However, as the species is not mentioned by Wells and Bryce to occur in Western Australia, I presume that its range does not include South Indonesia. Schilder also excluded the species from that area. Dharma's description obviously is not applicable for South Indonesia. Burgess' map should not include South Indonesia either.

C. quadrimaculata Gray 1824 was found only in Indonesia where it was common in Sulawesi and smaller neighboring islands. On East Java it was found a few times only, in West Java the species was not found at all. This demonstrates how the frequency of a species gradually decreases towards the limits of its range.

C. scurra Gmelin 1791 was found occasionally in the Seychelles and Indonesia. The specimens from the Seychelles had very little pigmentation. It was not found in Sri Lanka, but as the species is described as rare by Burgess, I can have missed it.

C. staphylea L. 1758 was found frequently in the Seychelles, Sri Lanka and Indonesia. The color of the shell is very variable. I found two main types: grayish blue and brown. In the Seychelles most specimens were poor in pigmentation, resulting in pinkish to yellowish shells with light yellow tips. According to Burgess, older and more callused specimens become milky white on the dorsum. I doubt whether this is age-related as I found creamy white specimens of small size in Indonesia which did not have much callus and could not be very old.

C. stolida L. 1758 was found occasionally in Indonesia and Sri Lanka. I did not find it in the Seychelles where it is said to occur occasionally as well. This fact, together with the observation that is was found regularly in East Java but not at all in Sulawesi, could be an indication for clustered distribution. Burgess' distribution map must be extended northwards over Sri Lanka. Also Schilder missed this distribution.

C. talpa L. 1758 was found occasionally in the Seychelles, Sri Lanka and Indonesia. Although an uncommon species, it was found frequently at some locations in West Java and was absent at others, another fact that hints at clustered distribution. Specimens from the Seychelles were significantly lighter in color.

C. teres Gmelin 1791 was found occasionally in the Seychelles, Sri Lanka and Indonesia. Especially in Indonesia its distribution is certainly not homogenous.

C. testudinaria L. 1758 was found occasionally on the Seychelles. In Indonesia I found it only in the Spermonde Archipelago, southeast of South Sulawesi. This indicates that its range extends further south than indicated on Burgess' distribution map.

C. tigris L. 1758 is probably the best known of all cowries, and certainly one of the most attractive. I found it in the Seychelles, Sri Lanka and Indonesia. Although literature still describes the species as abundant, frequent or common, this will be history in the not too far future, because the species is collected on a great scale for the tourist market. Specimens are seen in shops and on markets in large quantities, but on beaches hardly any can be found anymore.

C. ursellus Gmelin 1791. I consider this species only as a form of C. hirundo L. 1758 as in my opinion there are no conchological differences to justify species distinction. According to Burgess C. ursellus is smaller and with a contrasting dorsal pattern similar to C. kieneri Hidalgo 1906, but with fine teeth. As C. ursellus is reported to range from Japan to the Central Pacific and northern Australia, it is out of my range of observation.

C. vitellus L. 1758 was occasionally found in the Seychelles, Sri Lanka and Indonesia. The species varies widely in size with circular white spots on a brown background.

C, walkeri Sowerby 1832 was found neither in the Seychelles, where according to Slimming and Jarret it is rare, nor in Sri Lanka, which is outside its range. I only found it in East Java, were it is uncommon. The form surabajensis Schilder 1937 was found a few times in south Bali. According to Wells and Bryce the species is common in Western Australia, which indicate a strong gradient in frequency from east to west.

C. ziczac L. 1758 was found a few times in Indonesia. In the Seychelles I found the related C. diliculum Reeve 1845. I did not find it in Sri Lanka although according to Burgess that is in the species' range. But as its frequency is considered as rare, its absence in washed-up shell assemblages in Sri Lanka and the Seychelles is not surprising.

3. BEACH FOUND SHELL ASSEMBLAGES

3.1 Frequency lists of cowry species in washed-up shell assemblages

Washed-up assemblages of shells are the only representation of the living cowry population on the reef in front of the beach. However, these death-assemblages differ considerably from the living population as a result of transport and deposition. But as it is virtually impossible to study the living assemblages of cowries, as they are nocturnal animals that hide during daytime, the only way to perform quantitative studies is to use washed-up assemblages. Although it is unknown to what extend the composition of the living population is changed when washed-up shells are used, at least there is a certain constancy in it, as will be shown in this paragraph.

To obtain a frequency list of cowries I usually selected a strip of beach of about 100 meter long and started picking up all cowry-shells at the high-tide line. Here the larger species can be found. Next I collected closer to the water and continued until the low-tide line was reached. Here usually the smallest specimens are found. After collecting, usually for 3-5 hours, all material was identified as far as possible, and the number of specimens for each species was counted. This resulted in a frequency list, indicating the representation of species above a 4% minimum limit. Whenever possible, this procedure was repeated for a number of times.

3.2 Frequency lists of cowry species over a number of successive days at one location

During the period 1989-96, during July and August, I have been able to make frequency lists of cowry species washed-up on beaches at uninhabited locations with a steady wave action in Indonesia, Sri Lanka and the Seychelles. At three of these locations I had the opportunity to return a number of successive days, providing me the opportunity to investigate to what extend frequency lists of cowry species in washed-up shell assemblages are constant. The results are listed below.

Grajagan Bay, East Java, 1991

Nyang-Nyang, south Bali, 1991

Ranna, south Sri Lanka, 1996

Galle, south Sri Lanka, 1996

Provided the samples are large enough, the frequencies of washed-up cowry species remain fairly constant over a number of days. There are basic differences in the composition of beach-found thanatocoenoses at different locations, not necessarily far apart, like East Java and south Bali, and Ranna and Galle in south Sri Lanka. The question whether these differences are the result of processes like transport and deposition, or reflect differences between living populations, must remain unanswered.

3.3 Frequency lists of cowry species at annual intervals at one location

At four locations I was able to make frequency lists with an interval of one year in the month of July. The results are presented below.

Grajagan Bay, East Africa. Frequency lists for four years in July

Ujong Kulon, West Java. Frequency lists for two years in July

La Digue island, Seychelles. Frequency lists for two years in July

Galle, south Sri Lanka. Frequency lists for two years in July

These results invite the conclusion that the cowry thanatocoenosis is fairly constant, not only over a number of days, but also over a period of one or more years. Fact is that each location has its own characteristic frequency list for washed-up cowries. It is impossible to determine whether these differences are the result of transport and deposition, the occurrence of clustered distribution, or a different composition of the living cowry population.

3.4 Frequency lists of cowry species with a 6-month interval at one location

At one location in south Sri Lanka I had the opportunity to compare frequency lists of washed-up cowries in August and December. The results are presented below.

Ranna, south Sri Lanka. Frequency lists for August and December

All species keep the seem position in the frequency list, so at least for this location the frequency list is fairly constant both over a number of days and during the year.

3.5 Frequency lists of cowry species from different locations with similar conditions.

From the previous sections it is obvious that frequency lists of species of washed-up cowries remain fairly constant in time at one location. In this paragraph locations with similar geographic conditions will be compared. For this purpose four locations were chosen which resemble each other very much in longitude, geographic conditions and orientation towards the monsoonal winds. These locations are situated at Ponto and Genteng on West Java, about one hundred km apart from each other; and Nyang-Nyang beach on the south coast of Bali and Cape Aan on the south coast of Lombok, which are about 150 km apart and about 800 km from the locations on West Java. During my visits to these four locations in August 1991, all beaches were subject to heavy wave action, leaving the visitor in a continuous salt spray. The results are presented below.

It is obvious from these data that similar geographic conditions do not guarantee similar frequency lists for cowry species in washed-up shell assemblages at different locations with similar conditions. Where the small species C. kieneri and C. hirundo are most abundant in Ponto, West Java, their frequency drops to a fifth place in Genteng, West Java, and they do not occur in the lists of Bali and Lombok, meaning that their frequency is less than 5%. It actually was 1% in Bali and 2% in Lombok.

At the same time it was observed that C. isabella and C. helvola show the reverse: from being the most abundant species in Lombok with 14% C. isabella decreases to a mere 4% in Bali. Remarkable also are the figures for C. caputserpentis and C. lynx: both species are the most frequent ones in Genteng, West Java, decreasing from the 3rd and 4th place in Bali to the 5th and 6th place in Lombok, while both species were represented in Ponto, West Java with only 2%. This means that from one location the frequency of both species can decrease to the east as well as to the west.

To me, this seem to be arguments in favor of my theory of clustered distribution of species rather than species being distributed homogeneously over their range. Transport and deposition only are not likely to be the responsible factors for the unequal occurrence of species at locations with similar condition, as currents especially will favor the transport of shells according to their weight and shape rather than to their specific differences.

3.6 Frequency lists of cowry species washed-up on beaches with calm waters compared to beaches with heavy wave action

In this section washed-up shell assemblages found on beaches with calm waters will be compared to beaches with heavy wave action. Assemblages from La Digue island in the Seychelles were studied for this purpose. In July, one side of this small island faces heavy wave action, whereas the other side, only three km away, is surrounded by calm waters. The whole island is surrounded by a fringing reef. The results from a study in 1989 are presented below.

Comparing the lists, one can see that both C. annulus and C. caurica, being the most frequent species in washed-up assemblages on beaches with calm waters, are absent in assemblages on beaches with heavy wave action and replaced by C. helvola and C. caputserpentis. This means that almost half the total number of washed-up shells is replaced by a different species, respectively 52% and 42% on beaches with calm waters and beaches with heavy wave action. Moreover, C. isabella, with a frequency of 6% in assemblages on beaches with calm waters is replaced by C. asellus, with the same frequency of 6%. This means that a total of 58%, resp. 48% of all specimens of washed-up cowries is of a different species when beaches with calm waters are compared to those with heavy wave action.

A similar study was done in Indonesia in 1990. Washed-up shell assemblages on beaches with calm waters in Bali Barat Nat. Park in north Bali; on the north of Moyo island, situated to the north of Sumbawa; and beaches on the north of Salayar island, situated south of Sulawesi, were compared to the beaches of West Java, south Bali and Lombok, which are all exposed to heavy wave action. The results were given in paragraph 3.5.

At all locations, C. annulus is the most frequent species, representing 30-51% of all specimens. C. errones being the number two species in north Bali and Moyo is replaced by C. moneta on Salayar island. C. lynx and C. isabella were only found on north Bali and Moyo. C. kieneri/hirundo were not found on Moyo, while C. helvola was found in numbers on Salayar only.

If we compare these figures with the frequency lists of assemblages washed-up on beaches with heavy wave action, it is observed that three species found there do not occur on beaches with calm waters, namely C. arabica, C. carneola, and C. caputserpentis. At the same time three other species were found in numbers on beaches with calm waters only: C. kieneri/hirundo, C. erosa and C. errones. Actually, this is not a surprise as the first group found on beaches with heavy wave action have heavy shells that are unlikely being transported in calm water, where the second group, the ones from the beaches with calm waters, are either small species or do have light shells, making it likely to find them here.

Finally the frequency lists of three more locations are presented. They are from beaches of three small islands situated immediately east of Bali and called Ceningan, Lembongan and Penida respectively and made in 1990. Here the reef was seriously affected by human activity like growing agar. The results are presented below.

Surprisingly, the most common species on Ceningan, C. helvola, does not occur in the frequency lists of the other islands. The same is true for the number one species on Penida, C. lynx, which is not seen in the lists of the other islands. Species usually not mentioned in frequency lists because their frequency is less than 4% are more frequent here: C. nucleus with 8% on Ceningan, C. tigris with 8% on Lembongan, and C. arabica with 6% on Penida.
Due to the fact that beaches of these small islands are surrounded by complicated currents and the human activity on the reef, it is not surprising that the frequency lists of these islands show great differences from those made on undisturbed beaches.

3.7 Relation between biocoenosis and thanatocoenosis

The question how the beach found thanatocoenosis related to the living biocoenosis cannot be answered at the moment. In August 1993 I visited the reefs and beaches of Pulau Panaitan island in the Ujong Kulon Nat. Park in West Java in order to make an attempt to compare both two. The frequency list for about 2000 beach-found specimens was as follows:

By turning blocks and observing the grass beds, C. annulus was found to be the most abundant species with hundreds of specimens observed in a couple of hours. Dozens of specimens of C. moneta and C. lynx were found, while C. arabica, C. tigris and C. caputserpentis were found occasionally. So, in order of frequency the cowry biocoenosis reads as:

C. annulus
C. moneta
C. lynx
C. arabica
C. tigris
C. caputserpentis

It is obscure why the abundant species C. annulus and C. moneta are so poorly represented in the washed-up shell assemblages. The factors responsible for the transformation of the biocoenosis to the thanatocoenosis are not yet understood adequately.

So far, the following conclusions can be made about frequencies of cowry-species in washed-op assemblages.

1. Frequency lists of species of washed-up cowry-shells are fairly constant when result of successive days and lists with annual intervals are compared.

2. Locations with similar geographical conditions not necessarily show similar frequency lists for species of washed-up cowries.

3. Frequency lists of cowry species of washed-up assemblages of beaches with calm waters are markedly different from those with heavy wave action.

4. The frequency list of washed-up species of cowries is significantly different from the one of the living biocoenosis.

4. VARIATION IN BEACH-FOUND COWRIES

4.1 The growth pattern of cowries

In this section data on variation in characteristics like size, shape, banding, spots and pigmentation of beach-found cowry shells are presented. First an analysis of the growth pattern is given. Cowries, like all gastropods, stop growing in length when the outer lip is formed. What follows is thickening of the shell, resulting in widening and becoming heavier. Eight species of cowries were studied: C. arabica (N=45, from Java), C. boivinii (N=65, from Java), C. carneola (N=100, from Java), C. caurica (N=53, from the Seychelles), C. isabella (N=95, from Java and the Seychelles), C. lynx (N=63, from Java), and C. moneta (N=130, from Java and the Seychelles).

The shells were measured with calipers to the nearest 0.1 mm, and the weight was taken to the nearest 0.1 gram. A fixed ratio of 3/ 8 was found for width/ length for all species studied. C. moneta was the most variable in shape, the width of the shells varying between 3 and 5 mm.

In all species the weight/length ratio increases with the length of the shell, meaning that larger shells proportionally put on more weight than smaller ones. For C. carneola this weight/length ratio increases from 0.08 in small specimens to 0.16 in larger ones. On the average, weight increases from 2 grams for small specimens to over 4 grams for the larger ones. For C. moneta this ratio increases form 0.06 to 0.15 and occasionally more, with an increase of weight from 1 to 2 grams. All species showed this trend, reaching extreme values in C. caputserpentis: specimens of 22-24 mm weighing an average of 3 grams, increasing to 4.5 grams for specimens of 26-28 mm, and eventually weighing 7 grams in the 29-32 mm length category.

4.2 Origin of variation

It is a fact that members of a population of a living species share important features, but differ from one another in numerous ways, some rather obvious, some very subtle. This variation arises from different sources. Recombination of genes during sexual reproduction, together with crossing-over, the process in which parts of different parental chromosomes are exchanged during meiosis, results in the recombination of hereditary characters. Next, mutations in the genetic material can result in new alleles. Together these processes provide genetic variability on which natural selection can act to produce changes in a population. Moreover, there is phenotypic variability, produced by exposure to different environmental factors like temperature, availability of nutrients, etc. These variations do not have a genetic base and cannot be inherited.

4.3 Variation in size and shape

When washed-up mature shells of cowries from one location are measured and a graph is produced showing the size distribution of that species, one finds a basically symmetrical Gausse curve with the smallest specimens measuring about 50% of the size of the largest. However, a few exceptions were found. In C. carneola the curve being asymmetrical with a sharp drop on the left, indicating that few specimens smaller than the average size occurred in the population.

This phenomenon was very obvious in the shell assemblage of West Java and less pronounced in East Java shells. Another anomaly was found in the size distribution curves of C. isabella. Where the East Java curve is symmetrical indicating a normal size distribution, the West Java curve shows a second top, suggesting two forms of the species, a smaller and a larger one. SEE FIGURE.

After having measured as much as shells as possible of assemblages at different locations, the question arose to what extent the average length of adult shells of a cowry species is significantly different from the average length of the species measured at a different location. For 18 species of which sufficient numbers of shells were collected at different locations statistical analysis was done. In case the probability factor p equaled or was smaller than 0.05, the differences in average size between specimens of different locations was considered as significant, meaning that with a certainty of 95% or more the observed differences are not the result of chance, but must be contributed to a different factor. In doing so, four categories of cowry species can be distinguished:

1. Species not showing significant differences in size between locations at various places around the Indian Ocean.

2. Species showing a marked and significant increase in size in east-west direction over the Indian Ocean.

3. Species showing a marked and significant increase in size in west-east direction over the Indian Ocean.

4. Species showing a marked and significant increase in east-west as well as in west-east direction over the Indian Ocean when observed from a central location.

ad 1: Four species: C. arabica, C. carneola, C. helvola and C. kieneri did not show any significant jump in size in either direction when washed-up shell assemblages on beaches equally exposes to wave action on the Seychelles, Sri Lanka, East and West Java were compared.

ad 2: In eight species, a significant size increase in east-west direction was observed, not necessarily over great distances, e.g. C. ocellata. Specimens from south Sri Lanka (N=290) had an average length of 20.5 mm (s.d. 2.5= mm), a majority of 69% measuring between 17 and 22 mm. Specimens from south India (N=196) had an average length of 22.8 mm (s.d.=2.5 mm), and 79% measuring between 19 and 25 mm. See figure.

In assemblages from East and West Java there are significant size differences for C. caputserpentis, C. isabella, and C. lynx. For C. felina and C. interrupta significant size difference was observed between Java and Sri Lanka. Other species, like C. moneta and assumingly C. erosa show a size jump between Sri Lanka and the Seychelles. Summarizing, significant increases in size in east-west direction are observed over long distances as well over short ones. All size jumps are in the range of 6-13%.

ad 3: Four species: C. clandestina, C. gracilis, C. nucleus and C. staphylea showed a significant increase in size in west-east direction. Size jumps were observed between locations at various distances from each other. E.g. in C. staphylea a jump in size was found between East Java and south Bali, only 80 km apart. C. nucleus was significantly bigger in East Java compared to West Java, but there was no difference in size between West Java and the Seychelles. C. gracilis increases from Sri Lanka to West Java, while C. clandestina makes a jump of 19.2% in size from the Seychelles to Sri Lanka.

ad 4: For two species, C. asellus and C. annulus a marked and significant increase in size both in east-west and west-east direction was observed when viewed from a central position. For C. asellus there was an increase of 10.3% in east-west direction and 14.5% in West-east direction as seen from East Java. For C. annulus these figures were 11.4% and 10.7% respectively, also seen from East Java. All data are presented in the table below.

Species with significant differences in size between beach-found shell assemblages of different locations

Variation in shape

Besides size, also the shape of cowry-shells of one species varies considerably. Shape is expressed as the ratio width/length x 100. A low shape-index like 60 or less, indicates a slender shell, an index of 70 or more a rounder one. This difference of 10 points is easily seen. Again, considerable differences can be found between shell assemblages found not far apart, e.g. 76% of the specimens of C. ocellata from Sri Lanka have a shape index in the range 64-68; while 71% of specimens from south India fell in the range 62-65. This means that specimens from India tend to be more slender then shells from Sri Lanka. See figure.

For C. kieneri a marked difference in shape was observed between shells from the Seychelles and Java. In the Seychelles a majority of shells has an index in the range 51-55, in Java the index is in the range 56-60, meaning that shells from Java are markedly rounder than shells from the Seychelles.

C. asellus, C. carneola, C. gracilis and C. interrupta did not show marked differences in shape when different locations were compared.

Next, the question arises how to account for these differences in size and shape between washed-up shells from different localities. The size of a shell is not the expression of its age, because growth stops when the outer lip or labium is formed. What follows is only thickening of the shell. Since all shells measured did have a labium, the differences cannot be the result of age.

One might also think that the observed jumps in size are the result of local differences in transport and deposition, as it is in the unevenly distribution of left and right valves of bivalves. But if this was the explanation, it cannot be understood that other species apparently are not affected.

Another possibility is that the living cowries respond to favorable environmental factors by growing bigger compared to animals exposed to less favorable conditions. If a population of cowries is exposed to favorable environmental factors, one may expect that all shells will move into a range of higher values for length. One would expect the same response for all species, but this is not the fact as we have seen. Therefore, it is not likely that jumps in size are the result of environmental factors.

What remains is a genetic cause. If we assume that growth in cowries is controlled by a number of genes, as has been proven for several organisms, than this polygenic inheritance offers an explanation. If the dominant alleles affect growth in an additive fashion, than it is the actual number of dominant genes that determines the length of the animal. Although genes involved in multiple gene effect usually do not all contribute equally to the phenotype (length is this case), and the effects of so-called modifier genes and of the environment also tend to complicate the analysis of data, the idea of multigenic inheritance is very useful. If we also accept the presumption that some species of cowries are more apt to genetic changes than others, than it seems reasonable to assume that local populations can differ in size as a result of change in the allelic genes controlling growth. If in the course of time in some populations dominant alleles have changes into recessive one, a population will be smaller. A good support for this idea is the fact that size jumps have been observed in both west-east and east-west direction, as the loss of dominant alleles is supposed to occur at random.

Another argument in favor of this view is the observation that all jumps in size are of the same magnitude. As we have seen, these jumps vary from 6-15%, most likely the result of the fact that not all genes contribute equally to growth.

One might object that in this view the supposed genetic isolation is very unlikely when jumps in size between nearby locations like East Java and south Bali are compared. However, such isolation is not at all unlikely as those two locations are separated by very strong currents in Strait Bali. In the case of populations from the Adam's Bridge area in south India, and Sri Lanka, one must realize that monsoonal drift both in summer and winter results in currents passing underneath India and Sri Lanka, not penetrating the Gulf of Mannar. In this way the exchange of genetic material between populations is prevented.

Comparing populations in East and West Java, one must realize that the north coast and most of the south coast of Java are free from coral reefs, preventing an exchange of genes between the populations.

Whatever the final interpretation of the facts may be, the old idea of Schilder, stating that cowry-species grow larger from the center of their range towards the periphery, is definitely not true.

4.4 Variation in dorsal banding

Next, the variability in dorsal banding in cowries was examined. As length is a linear characteristic, and shape can be expressed as a simple ratio, I wondered how variability in dorsal banding could be expressed. Three species with marked dorsal banding were chosen for this purpose: C. asellus, C. carneola, and C. interrupta.

The total width of the bands was expressed by the ratio: sum of individual bandwidths/length of the shell x 100, measured over the dorsal ridge. The occurrence of wide or narrow bands was studies as well. A band was considered narrow or wide when it was at least 0.5 mm wider or narrower than any of the other bands

For C. asellus two assemblages of washed-up shells were studied: one from West Java (N=94) and one from the Seychelles (N=95). In both assemblages the total width of the tree dorsal bands varied from 39% to 69% of the length of the shell, there were no noticeable differences found between both locations. This means that the three dorsal band occupy between one third to two thirds of the shells length at both sides of the Indian Ocean. However, in the Seychelles only 42% of the shells studied had bands of equal width, versus 84% of the specimens from West Java.

For C. carneola, 42 shells from West Java were compared to 74 specimens from the Seychelles. At both locations the width of the double transverse dorsal band covered 24-39% of the length of the shell. The other characteristic measured in this species, the width of the interspace of the central transverse band in relation to the width of this band, ranged from 13% to 36% at both locations. The only difference being the absence of an interspace in 15% of the Seychelles' specimens.

For C. interrupta 36 shells from Java were compared to 27 specimens from Sri Lanka. The width of the dorsal bands and interspaces in relation to the length of the shell was in the range 64-74% for the shells from Sri Lanka versus 54-61% for specimens from Java. The width of the interspaces in relation tot the total width of the bands was about the same for both locations. At both locations and with equal frequency, either the middle band is the widest, or no band is wider than 0.5 mm or more than either of the others. For an interpretation of these data see 6.2: Genetic drift and loss of allelic genes.

4.5 Variation in oscillated spots

In this section attention will be paid to variability in dorsal spotting in those Indo-West Pacific species of cowries which are known for their oscillated spots. Three species available in sufficient numbers were chosen for this purpose: C. boivinii Kiener 1843 (60 specimens from East Java); C. gangranosa Dillwyn 1817 (40 specimens from East Java); and C. ocellata L. 1758 (over 200 specimens from Sri Lanka).

C. boivinii is characterized by Burgess (1985, p. 208) by having a slate gray or milky blue dorsum with indistinct brown spots surrounded first by a light zone and than a ring of darker brown. Lorenz and Hubert describe the species as: dorsum gray, profusely mottled with white, brown and oscillated spots. In the beach-found assemblage of East Java I noticed a gradual changing in spotting: from specimens with white spots and only a few brown ones, via specimens with an increasing number of brown spots and occasional oscillated ones, to a fully developed pattern of spotting with many small white spots, a fair number (20 or more) well developed oscillated spots, which are brown with a beautiful cobalt blue “eye” and up to 2 mm in diameter, and no brown spots, on a grayish to milky blue dorsum.

C. gangranosa is characterized by Burgess by 10-20 brown, often oscillated spots, and according to Lorenz and Hupert by small white spots and larger oscillated ones on a greenish-gray dorsum. Amongst the shells I collected in East Java, I observed a gradual changing from specimens with creamy white spots only, via specimens with an increasing number of brown spots among the creamy ones, to shells with a fully developed pattern of spots, including many creamy white ones of different sizes, together with brown spots and oscillated spots, which are brown with a light blue "eye" up to 2 mm in diameter.

C. ocellata is characterized by Burgess by having a tan dorsum with white to light tan spots, and oscillated spots with dark brown in their center. Lorenz and Hupert describe the species as having a brownish dorsum with oscillated black and white spots. Among the large number of shells I collected in Sri Lanka, a series of gradual transitions could be recognized. ranging from shells with many white spots and only a few oscillated ones, to specimens with many oscillated spots. Fully developed dorsal ornamentation included many white spots together with many oscillated ones of different sizes, the largest being up to 2 mm in diameter. For an interpretation of these data see 6.2: Genetic drift and loss of allelic genes.

4.6 Variation in pigmentation

In the Seychelles on beaches lining calm waters the phenomenon of poor pigmentation was observed in many species of cowries. For some species 80-90% of the beach-found shells showed pigment reduction. On beaches lining reefs exposed to heavy water action, the amount of shells affected by pigment reduction was no more than 10%.

Exsamples of pigment reduction in beach-found cowries in the Seychelles

The effect was observed in the following species: C. annulus; C. argus; C. asellus; C. caputserpentis; C. carneola; C. caurica; C. clandestina; C. cribraria; C. erosa; C. globulus; C. helvola; C. histrio; C. isabells; C. lynx; C. moneta; C. scurra; C. staphylea, C. talpa; C. teres; C. tigris; and C. vitellus.

Pigment reduction in shells was not found in Sri Lanka and it was occasionally seen in Indonesia. Wherever it occurred, specimens with normal pigmentation and those with reduced pigmentation ware found on the same beach, making it very unlikely to attribute this effect to deficits in nutrients. In chapter 6 on genetics this phenomenon will be discussed.

A different example of variation in pigmentation was seen in C. boivinii. In this species it is not the amount of pigment that varies, but the character of it. In 80 specimens from East Java I observed four types of color in washed-up shells: (1) with a grey dorsum; (2) with a milky-blue dorsum; (3) with a bluish-green dorsum; (4) with an olive-greenish dorsum. As seen in 4.5 there was also a lot of variation in the coloring of the spots: specimens with vague and blurred white, blue and brown spots were found among specimens with white and oscillated spots.

4.7 Intergrading variation between species

When large numbers of shells are studied, it becomes obvious that in a number of cases characteristics for species distinction are not constant, but highly variable instead. This, of course, has consequences for the taxonomy of the group. I will demonstrate this for a number of species.
C. kieneri Hidalgo 1906, C. hirundo L. 1758, and C. ursellus Gmelin 1791. The characters for distinction between the three species according to Burgess (1985) and Lorentz and Hubert (1993) are:

C. kieneri: (1) shape elongate-oval (L&H); (2) teeth rather coarse, extending except at anterior columellar side (L&H), teeth which are on the anterior columella markedly attenuated (Burgess); (3) base white, margins spotted (L&H); (4) dorsum covered with three bluish-green zones and small brown spots (L&H).

C. hirundo: (1) shape oval to cylindrical (L&H); (2) teeth fine, extending over base (L&H), teeth which are fine and cross most of the base (Burgess); (3) margins finely spotted (L&H); (4) dorsum with three blue to gray zones spotted with brown, often blotched (L&H).

C. ursellus: (1) inflated, with rostrate extremities and ribbed posterior (L&H); (2) teeth fine and reaching finely spotted margins (L&H); (3) dorsum with three dark bluish-green zones, mostly blotched (L&H). Burgess states that: "A study of literature leaves doubt about just what cowries are represented by the names C. ursellus and C. hirundo". Lorenz and Hubert however, are of the opinion that: "The fine but distinct ribbing of the posterior extremities [in C. ursellus] can easily be overlooked. There are no doubts on the validity of this name [C. ursellus] on account of this unique feature, despite its similarity to C. hirundo.

Careful study of 300 specimens of these three cowries I collected on Java, Indonesia, allows the following remarks. 119 shells could easily be recognized from the teeth as C. kieneri; 90 shells were identified by the fine teeth and posterior ribbing as C. ursellus. 78 shells with fine teeth but without posterior ribbing were taken for C. hirundo, leaving another 39 shells with fine teeth, but a very fine posterior ribbing or very barely visible lines in stead of ribbing. Length/width curves for these groups showed that each group of cowries had the same shape, making the description of C. kieneri as elongate to oval and C. hirundo as oval to cylindrical meaningless. However, these two species are distinguishable by their dentitions.

A more serious problem arises in separating C. ursellus from C. hirundo. As they have the same shape, the only conchological characters left for differentiating between these species are color and the presence of posterior ribbing in C. ursellus and the absence of it in C. hirundo. However, in the material I studied about 20% of the specimens with fine teeth have weak to barely visible ribbing or only faint lines on the posterior extremities. Therefore I have to conclude that this feature is not constant and hence useless for species distinction. That leaves the color of the shell, the dorsum with blue to gray zones in C. hirundo and with bluish-green zones in C. ursellus. But this is also not a clear and constant character. The animals differ only in the length of the papillae (shorter in ursellus), color of the tentacles (orange in hirundo, yellow in ursellus); and the presence (hirundo) versus absence (ursellus) of dust like spots on the siphon (Burgess, p. 156-157). Judging by the variability of these parts in other cowry species, these differences appear to be intra- rather than interspecific, a contention supported by C. hirundo neglecta Sowerby 1837, which is said to be anatomically identical with C. ursellus and possibly its synonym (Burgess, p. 157). As there appear to be no constant characters to differentiate between C. ursellus and C. hirundo, there is no ground in my opinion for regarding C. ursellus as a separate species.

C. histrio Gmelin 1791, and C. grayana Schilder 1930. Burgess (1985, p. 63) gives the following description of C. histrio: (1) spire blotch prominent and constant; (2) shell pale; (3) reticulations large, uniform and delicate; (4) barely noticeable longitudinal lines on dorsum; (5) base convex. C. grayana is described as: (1) spire blotch absent in about half, usually small and inconspicuous; (2) a definite hump on the dorsum; (3) longitudinal lines usually interrupted by the reticulations; (4) transverse embryonal bands very prominent; (5) base convex; (6) definitely pointed extremities in most specimens. Burgess concludes: "It (C. grayana) can usually differentiated from C. histrio by the darker transversely banded dorsum and the lack of a prominent spire blotch". Lorenz and Hubert (1993, p. 59) add to this: "oval to rhomboidal (C. grayana) and elongate to depressed (C. histrio); teeth fine, brownish (C. grayana) and teeth reddish (C. histrio)." And also: "The status of this taxon is nut fully agreed upon by all authors".

After careful examining 26 specimens of the C. histrio/grayana group collected on La Digue island in the Seychelles, I come to the following observations. (1) the characters for specific distinction between histrio and grayana: humped (grayana) versus not humped (grayana); and the absence (grayana) versus presence (histrio) of a prominent spire blotch are very variable. E.g. histrio-like forms with a prominent spire blotch occur with and without hump, and humped as well as not-humped specimens occur both with and without spire blotch. (2) transverse banding, characteristic for grayana, is equally present in humped and not-humped specimens and in specimens with or without spire blotch. (3) reticulations and longitudinal lines are not constant, neither is the color of the teeth. Taking all this into account, I cannot consider C. grayana as a valid species.

C. labrolineata Gaskoin 1849, and C. gangranosa Dillwyn 1817.Burgess (1985, p. 222-223) gives the following descriptions. C. labrolineata: (1) dorsum with circular discrete white spots, no dark spots of any kind; (2) no color in the canal. C. gangranosa: (1) dorsum with 10-20 discrete brown often oscillated spots; (2) color in the canal. Lorenz and Hubert (1993, p. 193, 202) add to this the following characters: C. labrolineata: (1) elongate; (2) teeth rather strong; (3) margins spotted; (4) extremities usually blotched; (5) dorsum greenish-brownish. C. gangranosa: (1) oval; (2) base and fine teeth white; (3) tips blotched; (4) terminal blotches distinct and divided; (5) dorsum greenish-gray. Despite all these distinctive characters they conclude: "labrolineata can be similar to gangranosa. Shells with spiny, dark stained teeth and others with shorter, plain white ones occur".

After careful examining 100 specimens of the labrolineata/gangranosa group, collected on Java, Bali and Lombok in Indonesia, I come to the following observations. (1) the main characters for distinction between the species: dorsum with white spots together with the absence of color in the canals (labrolineata) versus brown spots and the presence of color in the canals (gangranosa) are not constant, but very variable. (2) in about 10% of the "gangranosa" designated specimens, there are only 4-5 very faint brown spots instead of the 10-20 as indicated, together with very little color in the canals. Based on this facts, I cannot accept C. labrolineata and C. gangranosa as distinct. In my opinion this is intraspecific variation only.

C. cicercula L. 1758, C. globulus L. 1758, and C. bistronotata Schilder & Schilder 1937. Burgess (1985, p. 175) gives the following descriptions. C. cicercula: (1) spire blotch constant; (2) dorsum usually granulate; (3) basilar spotting absent; (4) extremities more produced than in any other of the group. C. globulus: (1) dorsum entirely smooth; (2) teeth short, anterior columellar teeth never cross the lateral carina; (3) sides and dorsum discretely spotted; (4) prominent posterior dorsal callus; (5) extremities less produced, blunt; (6) four basilar spots not constant. C. bistronotata: (1) dorsum granulate in adult specimens; (2) teeth more prominent, anterior columellar teeth reach the lateral part of the dorsum; (3) three pairs of brown blotches evenly spaced along the length of the dorsal sulcus; (4) four constant basilar spots. The main conchological character for distinction of C. bistronotata being the four constant basilar spots. Lorenz and Hubert (1993, p. 218) mention the absence of a spire blotch (bistronotata) versus the presence of it in cicercula and a humped dorsum without basal blotches in the Seychelles-form of globulus.

After examining 65 specimens of this group from the Seychelles and about 30 from Java, Indonesia, I come to the following observations. Among the Seychelles specimens with a smooth globulus-like dorsum there were five specimens with basilar spots, while the dorsum showed humps in different degrees of development. Among the specimens from Java there were two with well-developed bistronotata-like basal spots, but without a granulate dorsum. Taking into account Burgess' remarks about the living globulus in the Seychelles, which differs in coloration of the tentacles, siphon and foot from specimens in the Pacific region, I once more feel that C. gangranosa cannot be a valid species as the characters for its specific distinction are not constant.

C. pallida Gray 1824, and C. vredenburgi Schilder 1927. Burgess (1985, p. 162,163) says about these two species: "Cypraea pallida Gray is similar to both Cypraea vredenburgi Schilder and Cypraea xanthodon Sowerby. It can be differentiated from C. vredenburgiby the absence of a fossula and denticles". And also "The dorsum of C. vredenburgi is crossed by three brown embryonal bands". About the animal he remarks: "The animal (C. vredenburgi) resembles that of C. pallida. Lorenz and Hubert (1993, p.124) write about C. vredenburgi: "similar to pallida, clearly distinguishable only by its denticulate fossula".

After examining 30 specimens of this group from East Java, Indonesia, I noticed that all specimens had a fossula, and the embryonal bands being present in some specimens only and vaguely or absent in others. Also taking into account the resemblances of the animals of both types, there is, in my opinion, no sufficient reason to consider these types as distinct species.

C. carneola L. 1758, and C. leviathan Schilder & Schilder 1937. Conchologically the only character for differentiating between the two species is the margin of the shell, which is smooth in C. carneola or more or less tuberculate as in C. leviathan (Burgess 1985, p. 79). After examining 90 specimens from the Seychelles, where C. leviathan is not supposed to occur, I found full grown adult specimens with smooth margins among more or less tuberculated ones. So this only character for distinction appears to be very variable also. Both Burgess and Lorenz and Hubert mention a constant difference between the two types in the papillae: slender and flat with spear-shaped branched tips in C. carneola versus branched broad-based ones in C. leviathan, but is very questionable whether this is enough substantial difference for distinction on the specific level. In my opinion it is not and should C. leviathan be considered as a variety of C. carneola only.

5. QUANTITATIVE BEACH RESEARCH AND TAXONOMY

5.1 Introduction

The taxonomy of cowries is subject to different interpretations, a fact that is widely known among collectors, commercial dealers and biologists alike. At one hand are the so called "lumpers", considering all species of cowries confined to one genus, and accepting a certain range of variability without taxonomic consequences. At the other hand are the so called "splitters", breaking the cowry group into many genera, each with a large number of species and subspecies and varieties, each with a name of their own.

The result is chaos and confusion. Based on the modern biological concepts explaining the origin and purpose of variation, together with my results of measuring the amount of variation, I will try to explain why it does not make sense to designate small morphological differences to taxonomic levels. Only commercial classification is interested in an overload of names, because it makes collectors believe that those names stand for real and constant "subspecies" or "varieties".

5.2 Variation and the species concept

In previous chapters the term population has been used. In modern biology a population is defined as a group of individuals that interbreed and so share a common gene pool. Members of small populations resemble each other more closely than they resemble members of other local populations. This is the result of the fact that individuals of one population are more related genetically and are exposed to more similar environmental conditions. Hence, we may consider such local populations as temporary units that intergrade with other similar units .All local populations together make up the species.

Variation between local populations of a species is often correlated with geographic distribution. The farther apart two populations are, the smaller the chance of direct gene flow between them, and hence the likelihood that the differences between them will be more marked. Each local population tends to evolve adaptations to the specific environmental conditions in its own portion of the species range. Environmental conditions often vary in a more or less regular way with the geographical position, resulting in north-south or east-west gradients in most species. However, sometimes geographic correlated variation is not gradual, and there may be a rather abrupt shift in some character in a particular part of the species range. Such an abrupt shift in a genetically determined character sometimes is designated as a "subspecies", a term also applied to more isolated populations such as on different islands. But among biologists this term is not generally accepted. Many feel that the distinction is too often made arbitrarily on the basis of only one morphological characteristic.

Back to the species. It will be clear from the definition that the members of a species are distinguishable from each other in a variety of ways, but at the same time all members of a single species share certain important attributes and that, as a group, they are genetically separated from other such groups. Although the existence of discrete clusters of living organisms that can be called species has long been recognized, the concept of what a species is has changed many times in history. In modern views it certainly is not a static entity, typified by some ideal form as was the idea in the 18th century. As has been shown, it rather is a genetically distinct group of natural populations that share a common gene pool, reproductively isolated from other such groups. Of course, most species can be separated on basis of fairly obvious anatomical, physiological or behavioral characteristics, but the final criterion is reproduction. Morphological characteristics can only serve as clues towards the identification of reproductively isolated populations; they do not in themselves determine whether a population constitutes a species.

5.3 Systematics and nomenclature

It will be clear from the previous paragraphs that application of this theoretical definition is very difficult. Only one thing is sure: the view of classical taxonomists that personal judgment in deciding which characteristics should be considered and how they should be weighed in matters of nomenclature, is far too simple. Experience and subjective judgment are always involved, but the degree of subjectivity evident in classical taxonomy has motivated modern taxonomists to attempt to develop more objective methods. One of these new approaches is called phenetics or numerical taxonomy. It uses as many morphological characteristics as possible, weighs them all equally in the expectation that if enough characteristics are compared, the subjective judgment will be unnecessary. Other modern ways include molecular taxonomy, using the degree of hybridization of single-stranded DNA-molecules of specimens of two species; or the amino-acid sequences of proteins; or the DNA sequence itself.

Whatever method used, modern biologists try to group individuals into species, and species into genera in a way that indicates evolutionary history as well. In doing so, a genus is a group of related species, but nobody knows how close they are related. "Lumpers" like large genera and species with a certain range of variation without taxonomic designation, whereas "splitters" prefer small compact genera containing species that are very closely related. They also designate certain variations as "subspecies" and "varieties".

The current system of classification and nomenclature dates from the Swedish naturalist Carl Linné (Latinized Corolus Linnaeus), who lived from 1707 until 1778. His system already used kingdoms, classes, orders, genera and species. Phylum and family as categories were added later. As Linné worked a century before Charles Darwin, he had no conception of evolution and thought that each species represents an immutable entity. He was simply grouping organisms according to similarities, primarily morphological. Linné also introduced the binary nomenclature, the system of giving each species a name comprising of two words: first the name of the genus, and second a designation for that particular species.

Linné listed 42 species of the single genus Cypraea in 1758. About 60 years later more than 100 species were known and from that date the problems in taxonomy started. In 1884 Jousseaume recognized no less than 36 genera, a number that was reduced to 6 by Thiele in 1931. However, soon afterwards Schilder and his wife introduced no less than 4 subfamilies, 31 genera, 16 subgenera, only to be surpassed by Steadman and Cotton in 1946, who listed 13 subfamilies and 61 genera of cowries. Confusion was now complete. In the 1950's, Kay started the study of cowry anatomy in an attempt to correlate anatomy with classification. In the 1970's several authors reduced the number of valid species considerably. Burgess proposed one genus for all true cowries and recognized 202 species in 1985. In 1993, however, Lorenz and Hubert turned the clock back and re-introduced the taxonomy according to the German tradition as proposed by the Schilders, half a century before.

5.4 Splitters and lumpers

In this section the arguments of a lumper and of a splitter will be presented, followed by the results of quantitative beach research. A good example of a lumper is Burgess, whoe Cowries of the World (1970, 1985) is well known. His starting point is the observation that taxonomy and nomenclature of this group of mollusks is chaotic. He is of the opinion that due to the lack of other sources of information it is necessary to utilize conchological characteristics in most of the presentation of the species. He acknowledges extensive geographic variation, but in his opinion this does not justify taxonomic designation. Neither does he accept minor conchological characteristics to distinguish what he calls "so-called" species, as in many cases these variations appear to be inconsistent and impractical to apply, just like I have demonstrated in chapter 4 of this article. For these reasons Burgess only accepts 202 species of cowries and, at the same time, doubts the validity of quite a few. In his own words: "Conchological characteristics used in species determination should be consistent and of such magnitude that they can readily be seen and illustrated". From anatomical studies like done by Kay (1957, ' 59,' 64) he draws the conclusion that taxonomic grouping of the cowries into subfamilies, genera and subgenera as done by Schilder (1939, '41, ' 63), Cotton and Steadman (1946) and Allen (1956) has no anatomical basis and therefore is meaningless and misleading. This lack of anatomical differences between groups of cowries made him decide to revert to the use of the Linnean genus Cypraea for all cowries. As a last consequence of his view he is not afraid to recognize that: "the entire taxonomy of the Linnean genus Cypraea must be revised, as a result of which the species list materially may be reduced".

A good example of splitters are Lorenz and Hubert. They follow in their work A Guide to Worldwide Cowries (1993) the taxonomy by Schilder and Schilder (1971) in their Catalogue of Living and Fossil Cowries. The Schilders proposed division of the family Cypraeidae into 4 subfamilies, 11 tribes, 22 sections, 46 genera, 72 subgenera and 405 species. The only alteration Lorenz and Hubert made is the removal of the level of subgenus, a level they describe as clumsy. The justification for the endless lists of names is their opinion that: "Those with a specialized interest in cowries usually prefer a more differentiated scheme". And molluscan species in their opinion, can never be based on the rigorous definition of it, but rather on intuition and reasoning. They feel that whether cowries form a genus or a species, cannot be decided or disproved. The only condition a system must meet is that it must be compatible with the evolutionary process. It is from this premise that they introduce large genera with a small number of levels. But there is also a personal element involved when they state: "We feel that a structural treatment is, after all, to be preferred to the unstructural approach which is used by Burgess".

So, where Burgess demands clear, constant and distinct characters for distinction on the species level, Lorenz and Hubert are very subjective: "it is merely a suggestion based on our own observations". Based on this subjectivity they use the term subspecies as: "a segment of a species that is constant in general features separating it from the typical population". Based on their need for structural treatment, they also classify all variations and forms, but surprisingly enough do so "as a service to collectors".

5.5 One genus for all cowries

Finally I will bring up what contribution quantitative beach research can mean in this diversity of views. When splitters demand 'constant general features', even for designating subspecies, it is obvious that such features usually do not exist. Even characteristics used for species distinction are proven to be not constant, as I demonstrated for the histrio/grayana; labrolineata/gangranosa; pallida/vredenburgi and hirundo/ursellus groups. Of course, the definition of a species as a group of interbreeding individuals cannot be applied to cowries, but at least there must be a clear and constant difference between groups of individuals in order to designate them as different species. In many cases no anatomical, physiological or ethological information is available for taxonomic purposes, leaving only conchological characteristics to distinguish between groups. And such characteristics show natural variation to such an extend that it cannot be used for distinction. When authors indicate that their taxonomy must be compatible with the evolutionary process, such a view must be based on facts and not on subjective feelings.

Significant differences in size for different populations of the same species, which I observed in 12 species, are not recognized in the splitters view. This is not surprising as these differences were not known to them due to lack of specimens to be studied. Yet it is this information that contributes to a better understanding of the evolution of cowries. As long as so many questions remain unanswered, taxonomy must be kept as simple as possible. The consequence of this view is one Linnean genus for all cowries and the abolishing of using poorly described levels like subgenus and subspecies. If one regards taxonomy as part of the biological science, one has to take the consequences. Therefore, in my opinion, one can only be a lumper, leaving space for future knowledge, accepting incompleteness and not concerned with the interest of collectors and dealers.

6 POSSIBLE EXAMPLES OF GENETICS

6.1 Mendelian genetics

Genetics in gastropods is not a very popular topic. Compared to other living organisms, only a few examples are known. Maybe the best known example is the garden snail Cepaea nemoralis of which the inheritance of the banding and coloration of the shell are well studied. The basic color of the shell, which can be brown, pink or yellow, is determined by three allelic genes of which the one for yellow coloration is recessive. At the same time, dark bands can be present in different numbers. They can be separated or fused into one wide band. It is supposed that the presence or absence of these bands is determined by a single pair of genes.

Among the cowries an example of Mendelian genetics could be represented by the color of the dorsum in Cypraea boivinii, which can be milky blue, bluish green or olive green. These colors were found to be well distinct when studying specimens from East Java.

Another example could be the color of the dorsum in Cypraea ocellata from south India, which is either a yellow-orange to brownish orange, or olive blue.

6.2 Genetic drift and the loss of allelic genes

In section 4.4 it was stated that among specimens of C. asellus from the Seychelles there is much more variation in width of the dorsal bands than there is among specimens from Indonesia. Where 84% of the shells in Indonesia have bands equally wide, this is only 42% in the Seychelles. This could be the result of genetic drift, meaning a change in allelic frequencies independent of natural selection and occurring as a result of at random fluctuations in allelic frequency. Not only is natural selection in a stable marine environment unlikely, also the result can hardly be interpreted as an improved adaptation to the natural condition.

Another example might be represented by the variable dorsal spots in C. boivinii, C. gangranosa and C. ocellata. It is hard to understand how these spots can be a useful product of adaptation in an evolutionary process. We cannot imagine of any meaning of the presence of dorsal ornaments that are covered by a mantle in nocturnal animals. In other words, these dorsal spots cannot be imagined of as having a function in visual communication, neither between specimens of the same species, nor between individuals of different species. Speaking in evolutionary terms, we cannot believe an animal with a fully developed dorsal ornamentation in its shell having any profit from it compared with specimens with a less-developed pattern of complicated structures like oscillated spots. Therefore, it is unlikely that natural selection could have favored specimens with a more developed pattern of spots over specimens with less pronounced spots in the struggle for survival. If we adhere to the evolutionary view of life, we rather consider the well developed pattern of spots as remnants of a far distant era when different conditions prevailed. In this view, what we now observe is the result of a long period without any shaping force of natural selection on that particular character. Due to the lack of this force reduction of this character took place without affecting the survival result of the species. Gradual reduction of the dorsal pattern of ornamentation could then be explained as a result of the stepwise loss of allelic genes responsible for the full development of the patterns, resulting in a kind of dilution of the original genes in the gene pool of the species. This view is supported by my observation that we see the same thing happen in all three oscillated species in the Indo-West Pacific region.

Literature

Allen, J. 1956. Cowry shells of World Seas. Georgian House, Melbourne.

Burgess, C.M. 1985. Cowries of the World. Seacomber Publications, Cape Town.

Dharma, B. 1988. Indonesian Shells. Sarana Graha, Jakarta.

Kay, E.A. 1957. The genus Cypraea. Nature, vol. 180, p. 1436-37.

Kay, E.A. 1959. Generic Revision of the Cypraeinae. Proc. Mal. Soc. Lond., vol. 33, p. 279-287.

Kay, E.A. 1964. Anatomical notes on Cypraea aurantium Gmelin and other cowries and an examination of the subgenus Lyncia Troschel. J. Mal. Soc. Austr. , p. 47-61.

Krommenhoek, W. 1994. Beach-collecting cowries: Possibilities and Limitations. The Cowry n.s. 1, p. 17-22.

Krommenhoek, W. 1997. Twenty years of beach-collecting: a Survey. Of Sea and Shore, 20: 1, p. 45-49.

Krommenhoek, W. 1998. About size-differences in beach-found shell assemblages of 17 cowry species in the Indian Ocean region. Of Sea and Shore, 21: 3, p. 154-160.

Krommenhoek, W. 1999. Variability and Taxonomy in Cypraeidae. Of Sea and Shore, 21: 4, p. 221-226.

Lorenz, F. and Hubert, A. 1993. A Guide to worldwide Cowries. Christa Hemmen, Wiesbaden.

Schilder, F.A. 1939. Die Genera der Cypraeacea. Archiv fùr Molluskenkunde, 71, p. 165-201.

Schilder, F.A. 1941. Verwantschaft und Verbreitung der Cypraeacea. Archiv fùr Molluskenkunde, 73, p. 57-120.

Schilder, F.A. 1963. Lumpers and splitters. The Veliger, vol. 6, p. 104-110.

Schilder, F.A. and Schilder, M. 1971. A catalogue of living and fossil Cowries-Taxonomy and Bibliography of Triviacea and Cypraeacea. Mémoires Inst. Roy. Sc. Nat. de Belgigue, 85, p. 1-246.

Slimming, D. and Jarret, A. 1970. The Cowries of Seychelles. Phillips and Co., London.

Sorokin, Y.I. 1995. Coral Reef Ecology (2nd ed.). Ecological Studies, vol. 2. Springer, Berlin.

Wells, F.E. and Bryce, C.W. 1988. Seashells of Western Australia. West. Austr. Museum, Perth.