MacDougal: morphology of hepatic elateks. 53 



of the whole number were branched. Figs. 6, 7 and 10 were 

 from another sporogonium in the same head. Figs. 8 and 9 

 were from the youngest receptacle studied. Figs. 12 and 14 

 were from an older head on the same plant as that from which 

 figs. 8 and 9 were drawn. 



Although some of the sporogonia were fully ripe when 

 studied, in no instance was an elater observed which had lost 

 its membrane. 



Summary and conclusions. The above statements may 

 be summarized as follows: 



1. So far as known the young elaters of Liverworts always 

 contain starch. In most cases it disappears as the spiral bands 

 are formed, but it may also be i^resent in mature elaters. 

 Aside from this substance no cell-contents have been observed. 



2. The branching of elaters is known to take place in Tarfj- 

 ionia, Antlioceros, Radula and Conocephalus. It is probably more 

 general. 



3. In the case of Conocephalus, at least, the branching follows 

 a dichotomous order. It may be explained as due to the radial 

 arrangement of the elaters within the sporogonium, from which 

 the two ends of the elater come to have different capabilities. 



4. The conditions of branching may be said to depend upon 

 three circumstances: (1) The shape of the sporogonium, (2) 

 the arrangement of the elaters with reference to their mutual 

 pressure, (3) the structure of the elater. 



5. The branching does not take place until the pressure 

 within the sporogonium is relieved by the loosening up of the 

 spores and elaters, preparatoi'y to their being set free. There- 

 fore the branching is analogous to the phenomena of thyloses. 



6. The number of spiral bands in the walls of elaters varies 

 from one to five. These undergo branching and fusion. 



7. The normal elater of Conocephalus conicus, as a rule, con- 

 tains two spiral threads, one or both of which generally 

 branch. 



8. The abnormal branching of the elater causes abnormal 

 branching of the spiral threads 



The data on which this paper is based were gathered in the 

 morphological laboratory of the University of Minnesota. 



