278 



Crockett, Port Costa, and Martinez, as compared with shells from Goat Island, 

 Angel Island, Oakland Harbor, and the San Francisco waterfront. 



Coloration when present ordinarily manifests itself chiefly in a fairly broad brown- 

 ish stripe from dorsal to \-entral across the middle median portion. This stripe varies 

 in color from a pale brownish yellow through varying shades of brown, often warmly 

 tinged with red, to heavy sepia; the color nearly always tends to fade out toward the 

 dorsal and ventral margins of the shell. In addition to this median stripe we often 

 find, more especially in shells from the upper bay, a roseate or occasionally brownish 

 color suffused o^'er the denticulated area (anterior median) ; in the more heavily 

 pigmented shells a distinctly reddish or brownish patch occurs on the auricle. 



As regards the absolute differences of coloration in shells from the upper and 

 the middle bay, it is diiificult to make any statement more definite than that the 

 former manifest color in maximal, the latter in minimal, amount. Shells from the 

 upper bay are characteristically marked with reddish or brownish color, while those 

 from the middle bay are lighter and more translucent, and often entirely white. The 

 Dumbarton shells more nearly resemble, in this respect, those from the middle bay. 



With our present limited knowledge of the physiolog>' of Teredo, the basis of 

 these differences in color is difficult to determine. The color lies entirely in the peri- 

 ostracum, which may be scraped off, leaving the shell perfectly white. This would 

 suggest that the thickness of the periostracum is at least a factor in producing depth 

 of color. It has been stated above that the shells from the upper bay exhibit a thicker 

 and rougher epidermis; it is natural therefore that these shells should have the deepest 

 color, while those from the lower bay would be lighter, owing to the white prismatic 

 layer showing through the thinner epidermis. 



That coloration is, however, in considerable measure directly dependent on im- 

 mediate factors of environment, independent of the thickness of the periostracum, 

 becomes evident from certain types of what we may term "accidental" coloration 

 that have been occasionally observed. 



In two different instances we have found that a nail driven into a redwood timber 

 below water line has produced an area of black discoloration (ferric tannate) through 

 the action on the iron of the tannic acid in the wood. Teredos were working success- 

 fully in this discolored area, but their bodies and shells were deeply stained with blue 

 (fig. 112,6). Teredos working elsewhere in the same piece of wood were not so colored. 

 In a number of instances we have found that the shells of Teredo left in stagnant 

 water in our aquaria were colored almost entirely black by the action of sulphur 

 bacteria (fig. 112, 5). This must have been due to some physiological cause, as the 

 shells of dead teredos under the same circumstances were not discolored. 



X^-VRIATIONS IN THE PALLETS 



We have shown in figure 61, series of pallets from three localities in San Francisco 

 Bay, namely, Crockett, Dumbarton and Goat Island. In each of these series, the 

 pallet at the extreme left of the illustration is regarded as approaching the type of 

 navaJis, as figured by Jeffreys (1869) (see fig. 69). The other specimens in the series 

 represent departures in greater or less degree from the type. The smaller size of the 

 Dumbarton material is not regarded as significant; our specimens from this locality 

 were secured from timbers exposed only a few months, whereas specimens from the 

 other two localities were taken when older. But by the time the animal has reached 

 an age of one month the pallets have assumed normal proportions, and, so far as our 

 observations go, no changes ensue thereafter as a result of growth alone. 



