LINNEAN SOCIETY OF LONDON. XXXlll 



sions together in contmuous lines, as in Miquel's Flora Indo- 

 Batava. 



The recently published Flore Frangaise of GiUet and Magne 

 is, for species, almost exclusively confined to bracketed analytical 

 keys, into which more or less of descriptive matter is introduced. 

 This, accompanied by excellent typographical execution, seems at 

 first sight to offer great clearness and exactness ; but a closer ex- 

 amination creates some disappointment. The student does not 

 always want merely to be led to the name of his plant, but quite 

 as frequently to verify the name he has obtained by that or other 

 means ; and for this purpose these broken descriptions are very 

 inconvenient. And in point of execution, the endeavours to avoid 

 copying from De Candolle have not always been happy. Taking 

 the first few brackets of the general table, always of great import- 

 ance to the beginner, De Candolle had already in the second 

 bracket detached the large and easily defined groups of compound 

 flowers. This in Grillet and Magne's table is deferred to the sixth 

 bracket, — being preceded in the third bracket by the separation of 

 dioecious, from monoecious polygamous and hermaphrodite flowers 

 — ^a very puzzling one to the beginner, unless limited in the first 

 instance, as is done by De Candolle, to the separation of unisexual 

 from hermaphrodite flowers ; in the fourth bracket, by the dis- 

 tinction of tubular or verticUlate perianths from those reduced 

 to one or two scales, the student having here to include in tubular 

 perianths those numerous Cichoracese which he has to exclude 

 from them at p. 194 ; and in the' fifth bracket by the separation 

 of Papilionacese, a much less numerous and more complicated 

 group than that of compound flowers. The great rule to be kept 

 in view in these dichotomous analytical keys is, first, that the divi- 

 sions should be as clearly defined as possible ; secondly, that they 

 should divide the species as equally in numbers as possible. 



The tabular diagnoses in Leo Grrindon's ' British and Garden 

 Botany,' are evidently done with great care and well adapted for 

 popular use ; but the collateral subdivisions are sometimes much 

 too numerous. Class III. (p. 69), for instance, comprises seventeen, 

 which are far from being well contrasted with each other ; and the 

 utility of the book is in some measure marred by the disturbance 

 of the scientific arrangements, and a general appearance of disorder 

 in the details. 



In illustrated Floras, progress has been made both in Germany 

 and in our country. The new edition of our ' English Botany,' 

 in the full luxury of typography and illustration, is a great proof 



LINK. PEOc, — Session 1865-66. c 



