March 10, 1905.] 



SCIENCE. 



383 



throughout the germinal vesicle segregate 

 into an extremely fine spireme. This 

 spireme becomes shorter and thicker, and 

 shows a distinct longitudinal split. It 

 then divides transversely into eleven bi- 

 valent chromosomes, i. e., in each case two 

 univalent chromosomes remaining attached 

 end to end. There is no conjugation of 

 univalent chromosomes ; it is merely a ques- 

 tion of two univalent chromosomes already 

 attached in the spireme remaining so, this 

 causing the numerical reduction to half the 

 somatic number. 



The two univalent chromosomes of each 

 of the eleven bivalent chromosomes are still 

 attached end to end at the metaphase of 

 the spindle, and the longitudinal split seen 

 in the spireme of the earlier stage persists 

 until this period, causing typical tetrads, 

 i. e., bivalent chromosomes with a longi- 

 tudinal and a transverse furrow. These 

 chromosomes separate along the line of the 

 transverse furrow (which indicates the 

 point of attachment of two univalent chro- 

 mosomes). Allolobophora thus supports 

 the observations of Korschelt in the annelid 

 Ophryotrocha, and Montgomery and others 

 who maintain that the first division sepa- 

 rates univalent chromosomes and is, there- 

 fore, a reducing division. 



There are two nucleoli in each germinal 

 vesicle, the relatively large principal nu- 

 cleolus, and the smaller accessory nucleolus, 

 and neither appears at any time to be a 

 storehouse for the chromatin which forms 

 the chromosomes. The principal nucleolus 

 disappears before the spindle is formed, but 

 the accessory nucleolus may persist until 

 much later. 



We interpret the accessory nucleolus as 

 the precocious appearance of the nucleolus 

 of the oocyte of the second order. If, as 

 held by a number of investigators, the 

 chromosomes of one division are in some 

 manner related to the nucleolar substance 

 of the following rest stage, may not this 



take place at an earlier period, and the 

 accessory nucleoli of the germinal vesicle 

 be a precocious appearance of the nucleoli, 

 which are so conspicuously absent between 

 the first and second maturation spindles, 

 the processes involved in the rest stage oc- 

 curring before instead of after the first 

 division, and the origin and growth of the 

 accessory nucleolus being part of them? 

 The second division precociously foreshad- 

 owed in the four-part chromosome of the 

 germinal vesicle suggests a precedent for 

 this interpretation. 



A Quantitative Study of Holothuria atra 

 Jdger and the Beestahlishment of Holo- 

 thuria floridana Pourtales {=Holothuria 

 mexicana Ludwig) : Charles Lincoln 

 Edwards, Trinity College, Hai^tford, 

 Conn. 



Jager, 1833, described H. atra from 

 Celebes. Pourtales, 1851, described H. 

 floridana from Florida. Semper, 1868, 

 gave //. floridana as a synonym of H. atra 

 and since then all authors have followed 

 Semper. Ludwig, 1875, described H. mex- 

 icana. Clark, 1901, gave Porto Rican 

 specimens as H. mexicana. Ten of 

 these specimens identified by Clark are 

 included in my statistics and they are 

 H. floridana as defined in this paper. 

 Statistical analyses of 138 specimens, 20 

 from the Sandwich Island-Mozambique 

 and 118 from the Florida-Caribbean re- 

 gion clearly demonstrate that H. floridana 

 Pourtales should be reestablished as a 

 species distinct from H. atra Jager, and 

 that H. mexicana Ludwig is a synonym 

 of H. floridana. In general, biometry of- 

 fers a most important method in taxonomy 

 for determining the extent of variation and, 

 therefore, of the best (least variable) spe- 

 cific characters and their proper definition, 

 together with the separation of growth 

 from adult characters. The following com- 

 parison gives in brief resume the chief dif- 

 ferential characters: 



