964 FRESH-WATER BIOLOGY 



While it is not usually desirable to keep "dead" shells for the 

 cabinet, occasionally it is necessary. In such cases, the color can 

 be in some measure restored by applying a solution of parafiin 

 and gasoline (a cubic inch of the former in half a pint of the latter). 

 "Dead" Unionidae can be improved in appearance and the bril- 

 liancy of the nacre measurably restored by applying muriatic acid. 



A good collection is characterized by two essentials: 



First, the careful selection and preparation of the specimens 

 themselves; 



Second, absolute accuracy in the matter of the localities from 

 which they came. 



There is little excuse for having poor and ill-cleaned specimens. 

 There is none at aU for failure to keep accurate records of the col- 

 lector's fieldwork. A drawer of any common species, without any 

 indication of whence they came, even if well cleaned, would be 

 absolutely without value. Such a drawer of any of our species 

 from fifty or one hundred different localities, definitely indicated, 

 would be a valuable contribution to the conchology of any state. 

 Specific names can be supplied or corrected any time, but a mis- 

 taken or erroneous locality can never be corrected. The collector, 

 therefore, should be careful never to trust to memory for facts of 

 this kind. Both in collecting and cleaning, the specimens from 

 each locality should be kept carefully separated and labeled. Too 

 much importance cannot be given to this point. The study of the 

 geographical distribution of the mollusca is one of the most im- 

 portant branches of conchological work, and, to be of any value, 

 this must be based on absolutely accurate work on the part of the 

 collector. 



The maimer of casing and arranging the collection is largely one 

 of individual preference. A catalogue, however, is essential, and 

 it is better to begin systematically in this particular and thus avoid 

 the necessity of doing it all over again when the collection begins 

 to assume considerable size. There should be a serial catalogue 

 and a card catalogue. Each addition to the collection should be 

 numbered as soon as received and entered in the serial catalogue, 

 which should be a book ruled in as many columns as the coJector 

 desires. 



