From mid-1985 to the present (1991), this researcher has been engaged in a personal project involving some of the cowries found in waters along parts of the western and northern coasts of the island of Oahu in the Hawaiian Archipelago. Among the species collected, C. gaskoini represents a small (5%) but important portion of the more than 8,000 Hawaiian cowries collected during this period.

A large specimen of C. gaskoini, taken on the North Shore. It was on my 'farm' at Makua for 3 weeks to allow it to finish the dorsal pattern. mag.~2x – L=1.057 inches, Q=1.228

There are two separate underwater areas where the shells involved in this analysis were obtained. The larger is about the space of a major league baseball field, and is in the shallow waters off of Makua reef––far up Oahu's leeward coast. The depths here vary from 10 feet to 60 feet. The other area is about the size of that baseball diamond's infield, and is located on the northern shore of Oahu (North Shore). The depths here range from 40 to 60 feet. This site's exact location will not be revealed as others would readily exploit its commercially valuable populations.

This sample of C. gaskoini numbers 412 specimens. Of these, 184 shells (lot N) come from the North Shore, and the remainer come from Makua (lot W). All in lot N are live-collected as the bottom features in that area are not conducive to finding emptied specimens. The upper Waianae coast, however, provides much better opportunities to collect dead specimens, thus lot W is rather evenly split between dead- and live-collected shells. Freshly emptied shells are considered as live-collected.

An extreme specimen. Collected at Makua, it has the lowest Q value (most elongate shape) of any live-collected (or fresh-dead) C. gaskoini I have ever measured. L=0.533" Q=1.023 ~x2 mag.

These two lots have been divided into smaller groups along conchological lines and are defined (by me) as follows: Lot N becomes two groups; the first is labeled as "transitional" (a columellar callus is not part of the width measurement) and is signified by the mnemonic character _; second is the "moderately mature" group and its mnemonic character is m: shells in this second group possess a subtle columellar callus which affects the width measurement––a sharp, easily seen callus is extremely rare in this species. The lot from Makua (lot W) is similarly divided into both _ and m. Among the W sub-groups are dead- (B) and live-collected shells. No juvenile specimens are used in this analysis, nor are live- or dead-collected specimens taken by other collectors from other sites on either coast. Four damaged, distorted specimens from the North Shore are disallowed, also. This leaves 172 specimens in Lot N and 207 specimens in Lot W . The groups discussed will be the North Shore groups identified as m and _ plus the four Makua groups labeled as mw and _w (the live-collected shells), as well as mB and _B, the dead-collected specimens (I have no dead-collected C. gaskoini from the North Shore).

This analysis uses numerous line graphs in addition to its text and images to relate information. Such abstract data as length measurements and aspect ratios are not easily comprehended in large quantities. The graphs are a sort of "shadow" cast onto our reality by these vague entities. Fortunately, a researcher can––with a large, precisely measured sample––vary the position of the "light" source to reveal other facets of the data and garner more clues to its true nature. The lengths of the specimens were measured to the nearest 0.001 inch. These high resolution measurements and the large number of measured values reduce misleading maxima and minima in the plots. An additional set of length graphs––with its bins shifted 0.015" higher, slightly less than one third of the bin width of 0.051"––are also presented. These will be referenced in the text as the S plots (i.e., 1S, 2S, ...) while the primary set will be called the A plots. The S plots lose some of the features visible in the A plots but reveal other features, thus rendering a more complete picture of the data set.

Plots 1A & 1S; the Length values of C. gaskoini w/ (1A) an 'ideal' (Gaussian) curve (dotted line) for 379 shells, and (1S) Shifted bins plotting the same shells' lengths.

The analysis begins with the conjecture that Cypraea gaskoini Reeve, 1846, is a homogeneous species, meaning that it is the same from site to site (expecting some variation due to 'local' conditions). Plots 1A & 1S (at left; both Lots W & N are included: 379 shells) reveal a curve unrelated to the expected Gaussian distribution (bell curve), which does describe some cypreaid species in Hawaii. 1A has three modes (peaks); two major, and a lower, flat-topped minor mode. Included in plot 1A is a bell curve for a population of the same number of specimens and standard deviation. The bell curve's mode is at the mean of the sample. The baseline values are, from left to right, 3 standard deviations below the mean, the mean of this 379 specimen sample, and 3 standard deviations above the mean. The statistics for the sample are given in the APPENDIX. In plot 1S, the two major modes seen in 1A fuse into a platykurtic (flat-topped) major mode while the platykurtic minor mode of 1A resolves into two minor peaks. This species is not "normally distributed" in its L (length) parameter.

Plot 2A; the influences of the six sub-groups which comprise the L plot shown in 1A & Plot 2S; the same groups with Shifted bin origins.

In an attempt to find the cause(s) of such an abnormal graph, Plot 2A shows all six sub-groups (as described above) under the plot of 1A. So many lines can easily confuse the eye, thus the sub-groups will be displayed again, a few at a time. This hectic plot does allow one to get an idea of how each group influences the composite length plot. To understand the odd shape of the C. gaskoini length plot, the 379 specimen sample is fragmented into its constituent sub-groups. The results hold several surprises and suggest some rather interesting possibilities.

Plots 3A & 3S; North Shore specimens: all live-collected.

Lot N is displayed in plot 3A. The two sub-groups show a considerable degree of correspondence in both 3A and 3S with the exception that the m group does not produce as many longer specimens. The _ group has a plainly bi-modal shape. In plot 3S, the m curve resolves into three distinct peaks, two of which correspond with peaks on the _ curve. Also, the two group's plots correspond even better. These plots are remarkable for two reasons. First, the fact that two separate sets of length measurements follow each other so closely in two separate plots very strongly suggests that the peaks (and valleys) of the plots are true indications of the nature of the sample and not "noise" or a quirk of the graphing process. Second, the two groups must be very closely related.

Plots 4A & 4S; the Makua (Waianae coast) specimens: these are the live-collected shells.

Plot 4A examines the same types (m & _ ) from the Makua site. The most prominent feature here is the platykurtic peak produced by the _ group. This curve does not resemble its North Shore equivalent much in shape or position on the baseline. The m plot on graph 4A, beside being quite low in comparison, is also dissimilar to the much larger _ plot while the North Shore groups, in plots 3A & 3S, followed each other quite closely. The 4S plot substantiates the indications of 4A.

Plots 5A & 5S; live-collected m specimens from Makua and the North Shore: much in common.

In plots 5A (and 5S), I juxtapose the live-collected m groups from these two different locales. Immediately, one can see the correspondence of the Makua group to those specimens from the North Shore. In the previous plot (4A), it was plain that this small group did not fit well with the live-collected transitional ( _ ) specimens from their site at Makua. An obvious explanation is that the m population at Makua is coming from the North Shore population by way of veligers carried around Kaena Point in the prevailing tradewind-generated ocean currents. There is a strong implication in this graph that the North Shore veligers find the Makua 'dining' quite as good as that their which their stay-at-home counterparts have 'back home.' If these North Shore veligers find all that they need to grow to their typical size while browsing at Makua, what is keeping the bulk of other live-collected C. gaskoini at Makua (see 4A) from attaining similar lengths? I have collected both sizes from the walls of the 'Pocket' at Makua, as well as from the surrounding coral heads and ledges.

Plots 6A & 6S; the "transitional" ( _ ) specimens from both sites are plotted together.

In plot 6A, the live-collected groups of both sites are shown together for easy comparison. The two L plots have a minimum (near the mean) in common plus the lot W curve, as it falls off, has a peak near the main mode of the North Shore group and yet another peak coinciding with the other major mode of lot N 's curve. After viewing this arrangement, it is logical to surmise that some of the _ veligers have likewise been swept into the Makua area by those same currents and these also find these conditions acceptable.

Plots 7A & 7S; past and present populations at Makua suggest that something has changed.

Plot 7A illustrates the relationship between the dead- ( _B) and live-collected ( _w) specimens of the Makua site. While there are similarities, the dead (B) specimens produce a maximum that is very near to a normal distribution with a secondary mode protruding from the descending slope of the curve. The live _ specimen plot appears to be a "stretched" or elongated version of the dead specimen's curve, also with a secondary peak on its descending slope. This "stretching" is most probably due to the specimens of North Shore origin that had settled in Makua and were subsequently included in the Makua group. Of the seven mated pairs of C. gaskoini collected at the Makua site, not one of these was in the m group and only one of the fourteen specimens (0.722") exceeded the average length value of 0.671". This information, coupled with the indications of plot 5A, suggests that the North Shore individuals may not mate with the individuals of the Makua area. This may also mean that the callusing variety is not indigenous to the Makua area. Should this prove to be true, the ramifications would be of some consequence. More convincing data is needed to confirm the radical suggestion that two groups so similar may be ethologically separated and, thus, distinct species. At the very least they could be taken for distinct races of the same species.

Plot 8A shows the live-collected _ groups with the dead-collected mB of Makua. Plot 8S that same mB group is shown with the live-collected m groups of both sites.

The final length plot 8A has the live-collected _ groups and the mB group in it; 8S has all three m groups. 8A is actually plot 6A (live-collected, transitional specimens from both sites) with the dead-collected m group interposed. In 8A the curves produce an enigmatic graph as the dead-collected m sample reinforces the "notch" found in the other two curves. Why any particular length should be less likely to exist than any other probably has deep genetic significance because it is highly unlikely that the predators of C. gaskoini carry rulers to help them determine which specimens are not too short and not too long but just right. In 8S, the mB curve falls off before the main modes of the two live plots and does not not rise again as these others do. This could be another indication, if these dead specimens found at Makua are from the North Shore colony, that today's larger specimens of the North Shore are a relatively new occurrence.

Some additional information comes from 21 mated pairs of C. gaskoini taken on the North Shore. After comparisons, the occurrence of matched ( _ & _ or m & m) and crossmatched ( _ & m or m & _ ) couples appears to be Gaussian, or random. Ten of the 21 pairs were crossmatched while seven were of the m,m type and four were of the _ & _ type. Though the total number of pairs is limited, the observed arrangement here suggests that the individuals of these two groups are comfortable in whatever pairing occurs.

Plot 2Q: the bi-modal overall plot (in black) of the two C. gaskoini areas sampled. The dotted line represents the Gaussian 'ideal' for 379 shells.

Following is a discussion about what these graphs may have revealed: First of all, as stated above, the populations at these two sites certainly appear to be composed of differing groups of individuals. The living population in the Makua area is generally shorter than its North Shore parallel and the plot of its lengths is not obviously bi-modal in nature, whereas the longer North Shore population is clearly bi-modal and, seemingly, isolated from its relatives to the west. Reinforcing this concept of separateness are plots of a newly discovered (by me) parameter––here named the "shaping Quotient" (Q)––which confirm what was already visible to the eye, i.e., that there are two populations having clearly different shapes. The discovery of the Q parameter was the result of my attempts to combine the three physical measurements of the length (L), width (W), & height (H) of a cowry into a value that was not sensitive to the absolute size of the specimen. This value is produced by dividing the sum of the specimen's W and H (which are closely linked by growth during the bulla stage) by the L of that specimen. Values <1 (less than unity) are typical for elongated shells such as C. isabella or C. scurra while values >1 define inflated shells like C. tigris or C. cicercula.

Plot 3Q: North Shore population of C. gaskoini strongly tends to have an inflated shape. Plot 5Q shows that the live-collected m group of Makua seems to come from North Shore stock.

In reviewing the 2Q graph (above), it becomes clear that each site's population has a predominant shape. As can be seen in 3Q (at left), the North Shore group overwhelmingly favors a more inflated, globose shell. The live-collected _ specimens from Makua mostly stay to the lower side of the plot in 4Q (below) while the m population of Makua restates its close link to the North Shore m population here and yet again in plot 5Q.

Plot 6Q: a most curious plot of live-collected, un- callused gaskoini from Makua & the North Shore.

Next, in plot 6Q, the two live-collected _ groups produce a strange-looking "butterfly" pattern as each possesses a minor mode under the other's major mode. The minor mode on the Makua _ curve is most likely to have been caused by an influx of North Shore specimens and, indeed, I found that of the 15 specimens creating this minor mode, 10 are longer than the 0.671" mean. The complimentary minor mode of the North Shore _ curve is not as simple. A check of the specimens comprising that minor mode reveals that they are about evenly split on each side of the mean length value. Even though so many of the longer North Shore variety partake of the elongated profile more typical of the Makua population, there is still little cause to suspect that a regular "backwash"––or reverse current––brings the veligers from the shorter Makua population to the North Shore site. More is said about this oddly-patterned plot in the ADDENDUM.

Plot 4Q: the live-collected groups at Makua lack the close correspondence of their North Shore kin.

Because both types of North Shore specimens ( _ & m) come from the same tiny site and are found in intermingled pairs, the most obvious cause of their differences would be their dissimilar genetic composition. It is worthy of note that the m population of the North Shore site has no secondary peak in its Q plot, another indication of its distinctness. Plot 7Q (below) reinforces, once again, that the m group of shells from Makua are out of step with the predominating _ population found there. Lastly, plot 8Q (also below) seems to put the final nail in place as the Q plots for all m groups are juxtaposed. These three plots do not partake of the lower values typical of either the Makua group or the minor mode of the _ group from the North Shore. It seems reasonable to say that all callused specimens of C. gaskoini found at Makua have strong links to the population on the North Shore of Oahu. That population is clearly more inflated in shape than most of those currently found at Makua.

Plot 7Q shows the distinct shape of the Makua _ group, while 8Q shows how closely related, by shape, all m groups are.

It is not surprising that the ocean, wind, currents, and bottom features can team up to create an apparently effective barrier over such a relatively short distance. The geology of Yokohama Bay and Kaena Point combine to produce a sort of "backwater" for the tradewind produced currents. This apparently traps, or rescues, some of the North Shore veligers that might otherwise have been doomed to fall into very deep water. Indications given by the dead-collected groups of Makua, which are sort of a recent historical record of both sites, are that the North Shore population may be different today from what it was perhaps 50 years ago.

The dual modes of the _ North Shore specimens are quite probably another indication of genetic variances between this and the Makua population. Are these peaks due to genetic relics from the formation of C. gaskoini––who knows how many years ago––or is the North Shore population something that mushroomed since, or because of, Hurricane Iwa in 1982? The Makua specimens that represent the C. gaskoini currently living in Yokohama Bay also display secondary peaks but these are at, or near, corresponding peaks in the North Shore population and are, again, almost certainly due to the prevailing tradewind-produced currents strewing veligers downwind.

This researcher strongly suspects that the North Shore population of longer, inflated specimens of C. gaskoini is a relatively new occurrence, at least at that site. Due to the location of this particular North Shore site, it is unlikely that it was seeded from the Waianae coast, although the aforementioned hurricane caused untold changes to occur. The concordance of the m populations, in lengths and shapes, seems to be strong evidence that the tradewinds are responsible for this larger variety of C. gaskoini appearing on the upper Waianae coast. Additionally, the lack of any m type of specimen in a liaison with _ specimens of the Makua area, while not conclusive, tends to suggest that there may be an ethological (behavioral) difference in the two populations. This is especially curious when the North Shore population has been seen to have no difficulty producing mixed mated pairs. If––I repeat, IF––I were naming the two as apparent races of C. gaskoini it would be as follows: the North Shore population would called Cypraea gaskoini nui (Hawaiian for "large") while the Makua population would bear the name Cypraea gaskoini iki ("small").

In summing up this brief analysis, this researcher wishes to make it plain that these suggestions are by no means graven in stone. They are simply one thoughtful person's interpretation. They remain open for discussion and emendation. Nevertheless, it seems plain from the graphs of both the length and Q statistics that the nature of at least one species here in Hawaii is not so simple as many might have guessed (or hoped). Research into C. gaskoini, and other Hawaiian cowries, will continue. There are suggestions elsewhere in this researcher's findings that other species of cowries may vary from place to place and, probably, from island to island.

Plot 6Q: The minor mode of the North Shore population begs an explanation.

If both types ( _, m) of North Shore specimens began with the same shape, there should be similar, but distinct, Q curves produced by each type. Due to the callus on the m types, the m plot should be shifted slightly more toward higher values than the uncallused ( _ ) plot but, in fact, plot 3Q shows the reverse. The m population has its curve shifted toward slightly lower values. The specimens that are to develop a callus inherit, it would seem, a more elongated shape than their uncallusing siblings. As time goes by, the increasing callus alters its calculated Q value by increasing the width and thus the numerator in the formula Q=(W+H)/L. (The height is increased because the base accreted additional nacre as the columellar callus grew.) Any increase in the numerator, meaning in width and/or height, results in an increase in the value of Q for a given length. This causes the specimen's place in the plot of Q values to shift to the right, toward higher values.

Plot 3Q: the m group's plot would be further to the right than that of the _ group if both started with the same shape.

Therefore, the puzzling minor mode of the _n group of 6Q is caused, it can be conjectured, by the m specimens that did not get an opportunity to produce their callus (and thicker base) before being collected. This left them with their elongated shape and its lower Q value. It has been understood all along that the _ group must have some as yet immature m individuals passing through this stage on their way to acquiring their callused state, hence the name "transitional." Should it be determined that all of the shells in this enigmatic minor mode of the North Shore _ were destined to produce a callus, both groups would have nearly identical shapes but achieved by different mechanisms.

Normal callus development in C. granulata. L=1.017", W= 0.752", H=0.451" (~2x mag.)

This proposition of calluses––present, absent, or inchoate––has genetic ramifications which are not within the scope of this analysis. A reexamination of several of the subject specimens (m) shows that the columellar callus can account for as much as ten percent of the width of the specimen. The labial callus can approach a similar amount. These are modest percentages; a fully mature C. granulata may have increased its width as much as 50% due to labial and columellar callusing. In the image at the right, cuts have been made that stop just short of the inner cavity of this cowry (the lower cut and the upper, right cut). The black lines indicate where the true length and width measurements reside.

A cut into the labial callus of a _ C. gaskoini (~10x mag.)

The columella of the same cowry. L=0.955", W=0.645", H=0.493"

Similar cuts were also made into a "transitional" C. gaskoini (at left) from the North Shore and an m specimen from Makua (lower right), neither of which was included in the above study because their length measurements are compromised. When comparing the depth of the cuts in these shells, bear in mind the m specimen is half the size of the North Shore "transitional" shell. Like the C. granulata, these cuts were immediately adjacent to the line of the width measurement and stopped just short of penetrating the inner cavity.

A Makua mb shell: L=0.498", W=0.357", H=0.279" (~5x mag.)

Finally, not all plots of the various proportions of these cowries have been included since this is a brief analysis, however, plots of the width-to-height ratio continue along the same lines as those shown here in that they too indicate distinctive differences between the Makua and North Shore populations. These other differences are not mathematically linked to those discussed above. The slope of a linear regression between the Q parameter and the W:H parameter is essentially zero (i.e., R² <0.001).