TAOF'; SKOriSBKIM! 



also be observed witli regard to other characters in this sub-family, e. g, the number of filaments 

 on the sensory bristle of the fifth joint on the first antenna. — The bristles on the second and 

 third endite of the protopodite of the fifth limb showed complete constancy with regard to 

 luimber and almost with regard to type in all the species of the sub-family Cypridininae that 

 I had an opportunity of closely investigating. The bristles on the first endite of the protopodite 

 of this limb were with equal regularity subject to variation both in number and type. Curiously 

 enough the species of the sub -family Philomedinae that I investigated showed quite opposite 

 conditions. In these the bristles on the first endite of the above-mentioned limb were quite 

 constant in number and their ty^je too showed a rather marked constancy. On the other hand 

 the bristles on the second and third endites of this limb varied. — In this connection I may 

 point out as a curious fact the constant appearance of the three medial bristles inside the rostral 

 sinus of the shell in Doloria, Varguh; Macrocypridina, Hifhonostra, Cyp-idina (sensu str.) 

 and Cypridinodes, sub-genera which presumably constitute a natural unit within the sub-family 

 Cypridininae. The medial bristles situated near these three bristles are, on the other hand, 

 subject to considerable variation. This shows how in a rather large group details that are appa- 

 rently quite insignificant may remain constant. 



But it may be said by some that such small characters, such as bristles, etc., in which 

 the present work abounds, cannot, of course, be constant as a rule. With regard to this I wish 

 to state two facts. In cases in which I had an opportunity of carefully investigating a large 

 number of specimens of the same species, some hundreds for instance, as in the case of Philo- 

 medes glohosa (W. Lilljeborg). Pontocypris monsfrosa G. W. MtlLLER, some new species of 

 the genera Pontocypris, Xestoleberis and Krithe, I found such small characters sui-prisingh- 

 constant. (Of course no general conclusions are drawn from this.) In addition it is to be noted 

 that the variability of a quality in a species is to the classifier a fact of equal or almost equal 

 importance as the constancy of a character. If these small characters are not constant this 

 must be established. 

 .1 thorough revision \ATiat is Specially necessary under the present circumstances is of course a thorough 



. ' J ■'/'«"«« " ■ revision of the great majority of the species hitherto described. In doing this it would be best 

 he rnrrif'd out. to procccd vcry radically with all the forms that are not described so well that they can be 

 identified with complete certainty and of which it can be proved that there are no type specimens 

 in existence. Unless there are special reasons for not doing so, these species ought not to be 

 taken into consideration any further; it would be best to consider them as non-existent. It 

 does not seem right to devote a great amount of work to setting up more or less long lists of 

 synonyms, in which most of or sometimes almost all the names ought really to be followed by 

 a query; from a scientific point of view such lists do not seem to be any gain. 



In the same way it seems necessary to deal very radically with identifications made 

 from species that are described in an unsatisfactory way, whether they are nomina nuda or 

 unsatisfactorily re-described. In the first place these should not — unless, of course, there are 

 special reasons — be included in lists of synonyms and secondly they should not be used in 

 zoogeographical investigations if the specimen or specimens on which the statements in 

 question are based are not still in existence. 



