180 ACTINIADiE. 



character of their skin. He treated them accordingly, and 

 in a few days they assumed the usual plump condition. 



Facts seem to show that even the same individual is 

 liable to considerable change of colour. Mr. Cocks tells 

 me that from some hundreds of experiments he has ascer- 

 tained that " the colour is materially changed by diet, 

 good or bad ; by water, pure or impure ; by attention or 

 neglect ; by over- feeding or starvation." And Mr. E. L. 

 Williams, jun. has favoured me with still more precise 

 statements on this very species. He observes : — " A. me- 

 sembryanthemum does change. Bright green in two 

 months has got to dark olive in my tank ; bright amber 

 to dark brown ; brown with vertical yellow spots or dots 

 has lost these markings." 



Characteristic as are the marginal spherules, they are 

 subject to some irregularities. I found a large specimen 

 of the deep olive variety, which had on the exterior of the 

 margin two azure tubercles; — one of them round, well 

 defined, and in no respect distinguishable from the intra- 

 marginal spherules, — the other somewhat less so. Below 

 these, scattered down the side of the column, were four 

 or five more blue warts ; more irregular in form and 

 shape, but still well defined, and perfectly similar in their 

 azure hue to the spherules. I subsequently obtained a 

 second specimen with exactly the same peculiarities. On 

 the other hand, a specimen of the same variety — which 

 was sent me from Cumbrae by Mr. D. Bobertson about 

 six months ago, and is still in my possession — has never 

 showed the slightest trace of spherules, though in every 

 other respect perfectly normal ; the basal line and the 

 gonidial tubercles being of the usual azure hue.* It is 



* " M. Haime has remarked that these bourses chromatophores, or 

 calycine tubercles, are to the number of 18 in those individuals which 

 have not yet developed the tentacles of the 5th cycle ; of 24 in those 

 which have 5 or 5j cycles, and of 4S in those which have 6 cycles com- 



