208 CARNEGIE INSTITUTION OF AVASHINGTON. 



duplication, or fusion of elements, produce the markings of the different 

 genera, which in some cases are superficially quite unlike. In determining 

 the course of evolution of the color pattern of adults, the study of immature 

 specimens is of the greatest importance. 



In studying the color phases of some of the fishes whose power of color 

 change is greatest, it becomes very clear that the character of the changing 

 patterns in related species may be fixed with a degree of definitencss quite as 

 great as that which prevails when the distribution of pigments is more nearly 

 permanent. Scants punctulatus and croicensis have a mottled phase which, 

 considered in either alone, would seem to be the acme of irregularity. And yet 

 different though the pigments are, the color pattern is essentially the same 

 in the two cases. 



From the facts noted, a high degree of conservatism in color pattern may 

 be inferred, indicating that, if the present colors and patterns exist by virtue 

 of their selective value, their character is considerably influenced by the ante- 

 cedent pigments and patterns whose variations were the raw material upon 

 which selection operated. 



Additional facts concerning the colors and behavior of the reef fishes have 

 been noted which may not at present be concisely summarized. For example, 

 the labroid fishes Chloricthys bifasciatus and Iridio bivittatus, if kept in a tank 

 with loose coral sand upon the bottom, will bury themselves in it at dusk and 

 remain concealed until daybreak. The regularity with which the reaction 

 occurs and the close phylogenetic relation of the two species support the 

 belief that it is a natural response to the stimulation of light. The fish will 

 give the reaction at an}^ time during the day if light be excluded from their 

 tank. They are not visible upon the reef after dark, but I have not seen one 

 conceal itself in the sand there. The act of concealment requires only a frac- 

 tion of a second for its completion and therefore may commonly escape obser- 

 vation, though occurring regularly. Again, Chloricthys bifasciatus, nitidus, 

 and nitidissimus are apparently a single species. The differences between 

 nitidus and nitidissimus are transitory, the characteristic color of one at times 

 replacing that of the other almost instantaneously. The differences between 

 nitidus and bifasciatus are those existing between immature and mature forms. 

 Every appreciable degree of difference between the undifferentiated pattern 

 of ni'tidus and the highly specialized pattern of bifasciatus may be seen in a 

 single school of the fishes, together with the phase which was described as 

 nitidissimus. 



Researches at Murray Island, by F. A. Potts. 



The Rhizocephalan Cirriped Thylacoplethus Coutiere. 



This form was described by its discoverer from three infected specimens of 

 Alpheus collected in the neighborhood of Torres Straits, but as the material 

 was neither fresh nor well preserved, he was not able to give a sufficiently 

 full or correct account of the parasite. 



At Murray Island Thylacoplethus was found fairly commonly on a species 

 of Alpheus commensal with Comatida. Infected hosts show a large number 

 of small pink sacs attached to the thoracic legs and the abdominal swimmerets. 

 These sacs consist simply of a mass of reproductive cells or embryos sur- 

 rounded by a thin body-wall, and all the sacs are connected by a common 

 root system running directly above the ventral body-wall and sending off 

 delicate branches into all the appendages. The fine, colorless roots are only 

 to be detected in living or specially preserved animals, and it is on this account 

 that Coutiere considered a root system to be absent. 



