STRUCTURE OF FLOWERS. 59 



being, as it were, split into two, or divided, so' that in the 

 place of one stamen with a bilocular anther are found 

 two with unilocular ones : this is observed in one of the 

 stamens of Impatiens noli-tangere. 



I do not know that any anthers exist which have 

 naturally more than two cells, but this appearance 

 results in several cases from two causes : — 1st, Each cell 

 is most frequently divided into two small ones by a 

 longitudinal fold of the dorsal part, and this frequently 

 gives the anther the appearance of being four-celled : 

 2dly, It sometimes happens that two or more neigh- 

 bouring anthers are united together, so that the ap- 

 parently simple body which results, appears to be four-, 

 six-, or many-celled ; for example, in the Salix impro- 

 perly called Monandra, and probably in the Yew. We 

 shall revert to this subject on speaking of the Cohesion 

 of the Stamens. 



The cells open at maturity in four different ways : — 



1st. The most frequent case is that where the opening 

 takes place by a longitudinal fissure through the middle 

 of each cell ; and when the anthers are two-celled, we 

 say that they are Birimose. 



2d. It happens more rarely that the opening takes 

 place bv transverse fissures, as in the Lavender. 



3d. Several anthers open at their apex by two pores 

 (Biporose) situated at the top of the two cells — in 

 Solarium, for example ; or by one pore at the top of a 

 unilocular one, as in Amaranthus. This kind of dehiscence 

 may result from the longitudinal fissure remaining united 

 except at the apex. 



4th. The most singular case is that where the cells 

 open by valves which detach themselves from the base 

 upwards, as is seen in the Barberry and Laurel, and 

 in the two families to which these plants belong — the 

 Berberidac and Laurineae. 



