the present State of Botanical Classification. 49 



to be the nearest, though distant by a whole round of the 

 spiral. 



The extent of these alliances does not yet rest upon any fixed 

 principle. Linnaeus and Adanson throw plants into fifty-eight 

 assemblages, and we must consider a botanical family of that 

 day as representing the alliance of our own time. Bartling 

 makes sixty classes or alliances. The annexed table makes 

 sixty alliances also, averaging each about five families. Those 

 of Dr. Lindley are more numerous, averaging rather less than 

 three families; he seems to have looked for consolidation in 

 h : s groups, and precision in the alliances. The alliances of 

 the table sometimes correspond with Dr. Lindley's groups, as 

 in Curvembryosa?, Aggregosce, Glumosa?, Spadicosa?, but they 

 more generally correspond with his alliances; the smaller al- 

 liances of Dr. Lindley must however continue under any 

 scientific classification as suballiances. 



In the revised table, the improved nomenclature of Dr. 

 Lindley is of course usually adopted; his work on the Natu- 

 ral System must henceforth be a standard of reference every- 

 where; and there is much convenience in the definite manner 

 by which such reference points out the tribes and genera in- 

 tended to be included. His happy manner of designating 

 the alliances by a termination in ales is also adopted. In 

 amending the nomenclature, it were to be further wished that 

 botanists would regularly designate all the families from ge- 

 neric names, throwing aside names from species or obsolete 

 genera, and all characteristic or arbitrary or mutilated names. 

 Such a regular system might, by a slight modification of the 

 termination, lead hereafter to a convenient mode of indicating 

 the position of each family in the general system. 



The limits of families are often at present more a matter 

 of taste than of definite distinction: Arnott has in doubtful cases 

 made good use of suborders as superior to tribes and sections. 

 The average points to five families for each alliance, and it 

 may be as well to lean towards that number; nothing would 

 be more easy than to reduce the whole to that standard ; we 

 sometimes have the choice of more than one combination, or 

 sometimes a more easy adaptation to a ternary or senary 

 division. Numerical symmetry is, however, a suspicious cir- 

 cumstance in natural history, though it is possible that an 

 antecedent, normal, and succeeding state may occasionally be 

 distinctly marked. 



In the first approximations for forming the table, the cha- 

 racters were studiously kept out of sight. Every one must 

 feel, on trial, the bias against evidence, where it is required 

 to modify a scheme for the admission of something additional, 



Third Series. Vol. 11. No. 64% July 1837. H 



