NOTES 0\ THE IDENTIFICATION OF CH.ETOGNATHA. 69 



the coast of California in 1904. The method has been to sepa- 

 rate the material into three assortments, according as the preser- 

 vation was excellent, fair, or poor. A large series from each 

 assortment was examined in an attempt to isolate the taxonomic 

 characters, and the results from each individual specimen tabu- 

 lated. Proceeding on this basis I am in a position to utilize, in 

 all essentials, the characters Fowler ('05, '06) has chosen. In 

 the ensuing pages frequent use will be made of his reports. 



Methods of Preservation. 

 Killing has been tried with various combinations of acids, 

 alcohol, and formalin, but, of all reagents, weak formalin gives, 

 by far, the best results. Other reagents cause unequal con- 

 tractions, swellings, and gross distortions, thereby ruining the 

 material. Fowler ('05) advises killing separately in weak for- 

 malin then, after a short time, transferring them, for permanent 

 preservation, into from 5 to 10 per cent. On the whole this 

 method gives excellent results but I have found, in some cases, 

 that considerable curling results. Specimens in this condition 

 are difficult to handle and measurements are rendered more or 

 less inaccurate. The following method will overcome this curl- 

 ing and otherwise give most perfect results : Each Sagitta is 

 placed separately on a dry slide or cover-glass and allowed to 

 remain exposed to the air for several seconds. This causes the 

 animal to stick slightly to the slide. Then hold the slide in a 

 vertical position and apply formalin (5 to 10 per cent.) at the 

 upper end, allowing it to wash over the animal which will be 

 killed before it has time to loosen itself from the slide. Another 

 method of adding formalin, drop by drop, to a jar of Sagitta has 

 given excellent results, but, if the formalin is added too fast, the 

 results are uneven, owing to variability in the activity of the indi- 

 viduals, some contracting violently to a very weak solution while 

 others are apparently unaffected by it. The three methods herein 

 stated are all good ; the second takes the most time but produces 

 the surest results. 



Method of Measurement. 

 Fowler ('05) advises camera drawings as measurements other- 

 wise taken are frequently erroneous, in many cases not tallying 



