XXVlll INTRODUCTION. 



tion that has been ascribed to it, there is no doubt that insertion of 

 stamens does very often go along with essential differences of other 

 kinds; for example, it distinguishes with precision Rosaceae from 

 Ranunculaceee, Violacese from Passifloretc, Reaumuriea) from Nitra- 

 riaceee, Aurantiacete from Burseracece. But, on the other hand, 

 there is not only frequently, as may be well supposed, so shght a 

 degree of adhesion between the stamens and calyx as to render it 

 difficult to say whether the former are perigynous or hypogynous, as 

 in Galacineae, Tamariscineee, and many others ; but there are orders 

 which do really exhibit instances of both modes. Thus Eschscholtzia 

 has decidedly perigynous stamens, and yet it is undoubtedly a genus 

 of Papaveracese, the character of which is to have them hypogynous; 

 and all kinds of gradations, from the one form to the other, are 

 observable in Saxifrages). The stamens of Macrostylis, among the 

 hypogynous order Diosmese, are manifestly perigynous. In Gera- 

 niacese the genus Geranium has the stamens hypogynous, and 

 Pelargonium perigynous. Caryophylleae are arranged among genera 

 with hypogynous stamens, yet some of them (Larbrea and Adenarium) 

 are perigynous ; in Illecebreae part of the genera are perigynous, and 

 part hypogynous. The perigynous stamens of Turneraceae divide 

 them from Cistinese, to which they are closely allied. — The manner 

 in which the stamens cohere is sometimes an indication of affinity ; 

 for instance, they are monadelphous in Malvaceae and Meliaceae, 

 diadelphous in great numbers of Leguminosao, polyadelphous in 

 Hypericineae ; but more commonly this character is unimportant, as 

 in Malvaceae themselves, which have sometimes distinct stamens; 

 Leguminosse, which have very often such ; in Ternstromiaceae, which 

 have both united and disunited ones. — It not unfrequently occurs 

 that the conversion of the petals into stamens takes place imper- 

 fectly, in which case a part of the stamens are said to be sterile, and 

 this is sometimes a useful character for detecting affinities. Thus, in 

 many Biittneriacea; one-fifth are sterile and petaloid, in Galacineae 

 every other one, in Aquilarine;ic two-thirds, in Bignoniaceae the 

 uppermost of 5 is rudimentary. — A peculiarity of a similar nature is 

 the want of symmetry which sometimes exists between the petals or 

 sepals, and stamens. Supposing the flower to be formed without 

 abortion of any kind, and by a regular alternation of metamorphoses, 

 as is usually the case, the petals will be always some multiple of 

 the sepals, and the stamens of the petals ; and of course any irregu- 

 larity in this respect will destroy the supposed symmetry. This is 

 often a point of much importance to observe ; for example, in 

 Bora (Tinea; the stamens are always equal to the segments of the 

 corolla, and the flowers of that order arc consequently symmetrical; 

 in Labiata;, on the contrary, one at least of the stamens is constantly 



