434 



SCIENCE 



[N. S. Vol. XXIX. No. 741 



Vironment, as shown by photographB taken by a 

 submerged camera. 



The conspicuousness is of the sort typical of 

 warningly colored insects and is often associated 

 with formidable means of defense. 



The conspicuousness is not due to secondary 

 sexual coloration. 



These fishes do not show aggressive resem- 

 blance and such resemblance is unnecessary for 

 them, since their food consists chiefly of fixed 

 invertebrates. 



They do not show protective resemblance and 

 have no need of it, since the coral-reef habitat 

 affords them ample protection from their enemies. 



Their conspicuousness is not an instance of 

 warning coloration, since they are readily eaten 

 by the commonest piscivorous fish of the region 

 (Lutianus griseus), when removed from the reefs, 

 although this fish possesses color vision, forms 

 associations readily and retains these associations 

 for a considerable time, and has therefore the 

 qualities which would enable it to take advantage 

 of warning coloration in its food. 



The conspicuousness of the coral-reef fishes 

 has therefore not arisen through selection of any 

 sort, but is an expression of the action of internal 

 forces (race tendency), in the absence of coimter- 

 acting selection. 



The disagreeable qualities of warningly col- 

 ored insects is universally held to have been pres- 

 ent before these insects became conspicuous. They 

 therefore served at the start to inhibit the attacks 

 of vertebrate foes and thus rendered protective 

 coloration unnecessary for such insects. 



The nature of their food has rendered aggres- 

 sive coloration unnecessary to warningly colored 

 insects. 



The conspicuous colors of warningly colored 

 insects have therefore arisen in the absence of 

 selection, under immunity from selection. They 

 are to be attributed to the action of internal 

 forces unchecked by selection. 



Other conditions than inedibility may so limit 

 the attacks of vertebrate foes on insects (and 

 other animals) as to render them free from selec- 

 tion from this source and wherever, in such cases, 

 the nature of the food renders aggressive colora- 

 tion unnecessary the insects are immune from the 

 action of selection and free to develop conspicu- 

 ousness. Inaccessibility may thus condition con- 

 spicuousness, and probably does so in the case of 

 many edible butterflies. 



The theory of immunity/ coloration is proposed 

 as a substitute for the theory of warning colora- 



tion, while at the same time it covers certain 

 cases not covered by the theory of warning colora- 

 tion. Immimity coloration is defined as follows: 

 " Coloration, not sexually dimorphic, which ren- 

 ders an organism in its natural environment con- 

 spicuous to vertebrates; which has no selective 

 value, since it does not aid the organism in es- 

 caping vertebrate enemies by concealment (pro- 

 tective coloration), nor in approaching its accus- 

 tomed invertebrate prey (aggressive coloration), 

 and when associated with disagreeable qualities 

 is unnecessary as a warning to vertebrate foes of 

 the existence of such qualities (warning colora- 

 tion) ; it is conceived to have arisen through in- 

 ternal forces under immunity of the organism 

 from selection acting on its color characters." 

 The exclusion of sexually dimorphic characters 

 from the definition is provisional. 



The PartulcB of the Society Islands, and the 



Problems of Distribution and Isolation: H. E. 



Ceampton, Columbia University. 



The survey of the islands of the Society Group 

 of Polynesia was completed during the years 1907 

 and 1908, and the results have made it possible 

 to offer relatively final statements regarding the 

 variation and distribution of the species of Par- 

 tula that occur in the group. Each island pos- 

 sesses characteristic forms, that with two excep- 

 tions are absent from other islands. The two 

 peaks of Tahiti contain nearly similar forms; the 

 two separated halves of Huaheine have the same 

 species, although these exhibit more differences 

 than the species of Greater and Lesser Tahiti; 

 Tahaa and Raiatea are wider apart, although they 

 have the same encircling reef, and their species 

 are far more differentiated; finally, Borabora and 

 Moorea possess unique forms, in correspondence 

 with their total isolation from other islands. 



A comparison of the valley faunas in any and 

 all islands reveals a similar relation between 

 geographical separation and racial divergence, and 

 all the islands agree in demonstrating this corre- 

 spondence. Evidence was presented showing that 

 environmental influences can not be regarded as 

 the immediate factors for racial differentiation, 

 and that mutation has played a large if not ah 

 exclusive part in the process. The r5le of natural 

 selection is restricted to a purely negative part. 



The Experimental Modification and Control of the 

 Behavior of Characters in Crossing: W. L. 

 TowEE, University of Chicago. 



A Theory of the Modification and Origin of Char- 

 acters in Animals: W. L. Toweb, University ol 

 Chicago. 



