TRUE FORMATION OF THE GENERA. 123 



is not liable sometimes to a variation. Thus we find the pericar* 

 pium variable in Impatiens, Campanula, Primula, Papaver, 

 Ci>tus, Fumaria, and Arbutus ;... the calyx in Nymphjea, and 

 Counus; the corolla in Vaccinium, Convallaria, Andromeda, 

 Gentjana, and LiNUM;,..and the seeds in Ranunculus, and 

 Alls m a. 



If the flowers agree, but the fruits differ, the genus ought not 

 to be parted. Thus in those extensive genera, the Cassia, He- 

 dysarum, Sophora, Lavatera, Hibiscus, and Mimosa, so great 

 a number of species have been ranged under the same genus, on 

 account of the conformity in the flowers, though there is a varia- 

 tion in the fruit. 



That the figure of the flowers is more certain than that of th« 

 fruit, appears from many examples ; as from Campanula, Pri- 

 mula, Antirrhinum, Alisma, Hibiscus, Cistus, &c. ; but the 

 proportion, of the parts is subject to very great variation. 



The number of the parts is more liable to variation than their 

 figure, and is found sometimes to vary even upon the same plant; 

 as in Rut \, Chrysosplenium, Monotropa, Tetragonia, Euo- 

 nymus, Philadflphus, and Adoxa, in the flowers of all which 

 the number of the parts varies from" five to four. In these doubt- 

 ful cases, the natural number must be collected from the prima~ 

 ry flower ; but in the variations of the number of the parts, there 

 is a proportional affinity worth remarking. In flowers the sta~ 

 mina usually vary from ten to eight, and from five to four ; the 

 corolla and calyx from five to four, and the whole flower from 

 four to three; and ihe fruit also usually varies from five to three, 

 and from five to four. 



The situation of the parts is the most constant, very rarely 

 varying in the same genus. 



The regularity of the petals is not so much to be depended on 

 as some former botanists* have thought ; for we see in geranium 

 the European species have regular corolla?, but the African one.- 

 jrregular. 



* Ricinus in particular. AvxitOiL 



