382 



Dr. E. Pearl. A Biometrical 



[Nov. 15, 



to be in very early stages of the conjugation process. In the second series 

 were included only pairs which, on the same basis of nuclear condition, 

 were known to be about to separate shortly. Now if the observed homogamy 

 is a result of an equalisation of the members of the pairs in size during 

 the conjugation process itself, clearly the second of these two series ought 

 to show the higher homogamic correlations, because the individual pairs 

 are in late stages, and we should expect to get the maximum effect of any 

 equalisation which had occurred. As a matter of fact the coefficients of 

 direct homogamy in length are - 6797±0'0307 for the " early " series and 

 - 6212 + - 0334 for the "late" series. The coefficients are sensibly equal 

 for the two, with what advantage there is in favour of the " early " rather 

 than the " late " series. From this it is concluded that the observed 

 homogamy is not due to an equalisation in size during the conjugation 

 process itself. 



8. From evidence which is discussed in detail in the complete paper, it 

 appears probable that the observed homogamic correlations arise as the 

 result of a process, the essential factors in which may be outlined as follows : 

 («■) At the periods of conjugation the individuals which are to conjugate 

 are in a so-called " miscible " condition in which the oral surfaces are 

 adhesive ; (li) as a result of the action of the currents, produced by the oral 

 groove cilia, two individuals which by chance happen to be swimming 

 parallel and close to one another are drawn together and their oral surfaces 

 adhere in whole or in part ; (c) the extreme anterior ends of the oral grooves 

 firmly adhere to one another first ; (d) if the two individuals are so nearly 

 the same size that the mouths approximately coincide when the anterior 

 ends are together, firm union occurs at the mouth regions, and definite 

 conjugation follows ; (e) if, on the other hand, the mouths do not approxi- 

 mately coincide, the individuals separate again or die and no conjugation 

 results ; (/) the homogamic correlations arise then as a result of the 

 necessity for the mouths of the two individuals to come together (or " fit ") 

 when the extreme anterior ends are united. Individuals in which the 

 distances from the anterior end to the mouth are approximately equal, will 

 not be greatly different in total length, and hence their lengths will be 

 correlated. The direct homogamic correlations for breadth and index, and 

 the cross homogamic correlations, are held on this view to have arisen 

 because of the organic correlation between length and breadth in the 

 individual. 



9. In concluding, some of the theoretical bearings of the work are 

 discussed. It is pointed out that if the results obtained in this work with 

 reference to the existence of a differentiated " conjugant type " are true, they 



