666 



SCIENCE. 



[N. S. Vor,. XXIII. No. 591. 



upper molars, and as conclusive evidence of 

 the correctness of this view has cited the upper 

 molars of Triconodon, in which the main 

 cone is central, and Dryqlestes and other 

 forms, in which he states the main cone is 

 internal. 



Owing doubtless to the incompleteness and 

 minuteness of the teeth examined, Osbom was 

 apparently led to error in observations, for 

 instead of three cusps, one internal and two 

 external, as stated by Osborn, the upper molars 

 of Dryolestes have five distinct cusps, one in- 

 ternal, three external and one posterior 

 median. This arrangement admits of a dif- 

 ferent interpretation of the cusp homologies. 

 The three outer cusps supported by two fangs 

 now appear to be homologous with the three 

 main cusps and two fangs of Triconodon, the 

 inner cusp being readily interpreted as a sec- 

 ondary or internal heel development. This 

 view is strengthened by a third type, Dicrocy- 

 nodon, in which the outer portion of the upper 

 molar is very similar to that of Dryolestes, but 

 the large inner cusp is totally different. In 

 Triconodon a broken external cingulum and 

 two incipient inner heel-like cusps preclude 

 the probability, at least, of this form of molar 

 ever passing to a typical triconodont stage 

 through the outward shifting of its lateral 

 cusps. These forms, therefore, apparently 

 represent distinct types of molars separately 

 derived from the simple cone, and Triconodon 

 and Dryolestes do not represent successive 

 stages in the development of the trituberculate 

 molar, as supposed by Osborn. 



Thus, the evidence of the Jurassic mammals 

 apparently agrees with the embryological evi- 

 dence and supports the ' premolar-analogy ' 

 theory, while it lends no support to the tri- 

 tubercular theory in so far as it involves the 

 position of the primary cone. 



The second paper was by Mr. M. 0. Marsh, 

 on ' Hemoglobin Estimates and Blood Counts 

 in Fishes in Health and Disease.' The species 

 observed were the brook trout and rainbow 

 trout. Apparently normal brook trout from 

 the Au Sable Eiver in Michigan varied widely 

 in hemoglobin and no norm was established 

 save a very broad one. Thirty-five specimens 



gave an average reading of 43 with the Dare 

 hemoglobinometer, 100 representing normal 

 human blood. The readings varied from 33 to 

 59. The hemoglobin of domesticated brook 

 trout averaged 34 from 23 observations. Wild 

 rainbow trout, represented by only two closely 

 agTeeing readings, were 92 in hemoglobin, 

 while the same species domesticated averaged 

 54 from 19 observations. The chief con- 

 clusions of interest derived from these figures 

 are that trout blood is lower in hemoglobin 

 than human blood, that the brook trout, 

 whether wild or domesticated, is considerably 

 lower than the rainbow in the corresponding 

 condition, and that domestication is attended 

 with a considerable falling off in hemoglobia 

 content. This latter fact is possibly corre- 

 lated with the increased susceptibility to dis- 

 ease in domesticated fish. Such a correlation 

 is more forcibly suggested by the comparative 

 insusceptibility of rainbow trout, even in 

 domestication, to a bacterial disease which un- 

 der exactly the same conditions attacks readily 

 the brook trout. The question is raised of 

 the possibility of increasing the hemoglobin 

 in the blood of the latter species by feeding 

 iron salts, as in human medicine. 



The wild brook trout has about one million 

 red corpuscles per cubic millimeter and the 

 number is not diminished in domestication. 

 The rainbow trout domesticated has 1,487,000, 

 being the average of eleven individuals, while 

 a single observation of the wild rainbow 

 showed 1,830,000. 



Trout in fish-cultural ponds have occa- 

 sionally true neoplasms of a malignant nature 

 located in the region of the gills and causing 

 a pronounced anemia. Ten brook trout thus 

 afflicted had an average hemoglobin reading 

 of 17. An apparently primary anemia in the 

 young of this species has been observed, in 

 some individuals so extreme that the gills in 

 life were white. The red cells had fallen oS 

 greatly in number, the lowest count recorded 

 being 38,000. 



On the other hand, the most destructive 

 epidemics of protozoan and bacterial infection 

 in trout are not attended by any marked 

 anemia. 



