162 CARNEGIE INSTITUTION OF WASHINGTON. 



ones were of two kinds. The first of these had dark bands that were broader 

 than those on the fishes in the preceding experiments, while those of the 

 second had them so much narrower that they were not as much darkened by 

 them as the striped fish were by their markings. In tests with colonies which 

 had had no previous experience in the matter, when the snappers were fed 

 first a stripe, then a narrow band, a stripe, a broad band, and so on indefinitely, 

 it appeared that stripes were still accepted more readily than bands, suggesting 

 that there is something about the banded pattern itself, as opposed to the 

 stripe, that causes the snapper to hesitate longer before accepting it. But 

 the experiment should be repeated. 



Plain, banded, and striped fishes are taken at first at different rates, but, as 

 the feedings are repeated from day to day, the differences in rate disappear. 

 When, then, in two colonies, one at the laboratory dock and the other under 

 one of the docks at Fort Jefferson, the point had been reached where plain, 

 striped, and banded fishes were being taken at the same rates, the established 

 procedure was changed. 



At the laboratory dock the striped fishes were made distasteful by sewing 

 into their mouths bits of the tentacles of the medusa Cassiopea, while at the 

 fort it was the banded that were so treated. Without attempting to enter 

 into details, it may be stated that after a time the snappers at the laboratory 

 dock showed a marked preference for the banded fish, and those at the fort for 

 the striped, even when neither of the two sorts offered to a single colony had 

 been made distasteful through use of the medusa tentacles. It may therefore 

 be accepted tentatively that the snappers are able to discriminate between the 

 banded and striped food-fishes offered them and to associate the distasteful- 

 ness mentioned with either of the patterns. Since the reaction once estab- 

 lished remained unchanged in both cases at a later time of testing, it may be 

 added that the snappers not only discriminate between the two patterns, and 

 form associations between either of them and qualities of the fishes displaying 

 them, but the association persists for a longer or shorter time. 



It is not possible at this time to say how often a colony of 50 or 100 snappers 

 must suffer from the Cassiopea sting before it avoids the fishes which carry it, 

 for it is highly probable that in my early experiments the stings were not 

 active in all the atherinas. A strong suspicion that such was the case led to an 

 investigation of the viability of Cassiopea fragments under the experimental 

 conditions. As one result of this investigation it appeared that when 125 

 atherinas of medium size have been preserved in 1 per cent formalin for some 

 time it is not sufficient to rinse and soak them in a large quantity of sea- 

 water, even for 2^ hours, before proceeding to use them. For when such 

 fishes, at the end of 2)4 hours, are again rinsed, soaked, and put in sufficient 

 fresh sea-water to cover them, and this smaller amount is twice or thrice 

 changed at intervals of half an hour, the extract continues to give positive 

 results when tested for formalin and kills jellyfish in a short time. 



Therefore I feel justified in expressing my conviction that Professor Reig- 

 hard's experiments are not conclusive. It is highly probable that the Cas- 

 siopea fragments which he believed were still alive, but did not test to de- 

 termine whether such was the case, had had their effectiveness greatly 

 reduced or destroyed by formalin extracted from the fishes in which they 

 were sewed. This conviction is supported further by the fact that his 



