130 



ther. At the basis the entire bract was indeed somewhat con- 

 tracted, but not so much so that this part might be called a 

 true filament. In those bracts which approached still more to 

 normal anthers, (and this was the case in most of them) there 

 were upon the outer side two protuberances, containing pollen, 

 of a long oval form and yellow colour, which at their hinder 

 end were partially separated from the bracts as in Araucaria. 

 These formations differed from the normal anthers only from 

 their being smaller, possessing a larger band and exhibiting a 

 total want of the filament. A portion of these anthers burst 

 longitudinally in the sutures, scattered the pollen, and then 

 dried ; others on the contrary remained closed and kept fresh 

 for several weeks, thus again exhibiting their nature as bracts. 

 From the phaenomena which the examination of these her- 

 maphrodite catkins presented M. Mohl deduces the following 

 conclusions : 



1. Each bilocular anther ofPinus and the allied genera has 

 proceeded from the metamorphosis of a single leaf, which is 

 also the opinion of Robert Brown and Lindley. 



2. The anthers of Pinus originate from leaves that belong to 

 the male catkin, and are not, as Mr. Lindley supposed, to be 

 considered as lateral monandrous flowers, and consequently 

 also are not analogous to the carpel-leaves of the female 

 catkin. 



Now it appears therefore necessary that the multilocular 

 anthers of other genera of Conifer should also be considered 

 to have originated from the metamorphosis of a single leaf. 

 M. Mohl is quite right in stating that the presence of a greater 

 number of anther-loculaments cannot be advanced as a reason 

 against this view, for it must not be overlooked that there is no 

 reason why pollen could not be formed in the interior of a leaf 

 as well at twenty different places as at one, at two, or at four. 

 Moreover the male amentum of Juniperus presents in the 

 number of anther-loculaments real transitions to various other 

 genera of Coniferce. For instance, the inferior scales of the 

 amentum of Juniperus very frequently possess only two locu- 

 laments, in which they coincide with Pinus ; the trilocular are 

 similar to those of Cunninghamia, and form the transition to 

 the quadri- and multilocular anthers of other genera. M. 

 Mohl here also touches on the explanation of the multilocular 



