188 MR. G. BENTHAM ON EUPHOEBIACEiE. 



bulk of the Order was undertaken by Jean Mueller of Argau, 

 who had then the charge of De Candolle's herbarium. He 

 devoted several years to this arduous task, visiting the herbaria 

 of Paris and of Kew, and published the result as the chief por- 

 tion of the same second part of the fifteenth volume of the 'Pro- 

 dromus' in the year 18G6. This work, as well as Baillon's, 

 showed great pains taken throughout to ensure accuracy of obser- 

 vation ; and if any defects are to be sought for, they are the very 

 reverse of those observed in the case of Baillon. There is no 

 want of divisions, subdivisions, and diagnostic conspectuses, but 

 their practical utility is much marred by the strict adhesion to 

 favourite characters, often exceedingly difficult to observe or even 

 purely theoretical (as in the difference between suppressed and 

 deficient petals), to the very general neglect of natural or geo- 

 graphical affinities. Geographical distribution may, indeed, have 

 been rather more taken into consideration by Mueller than by 

 Baillon ; but both authors appear to have too frequently allowed 

 natural affinities to be overruled by isolated characters to which 

 the one or the other had attributed a constantly prominent value : 

 as, for instance, where the one unites Seidelia with Tragia, far 

 away from its closest allies Adenocline and other Mercuralioid 

 genera ; whilst the other removes Calycopeplus and Anthostema 

 from EupTiorhia to place them in a far distant series, the one be- 

 tween Amperea and Cnesmone in his Jatrophece, the other next to 

 Dalembertia at the end of his Excoecariece. Neither of these 

 botanists appears to have borne sufficiently in mind the fact that 

 characters differ in value in different genera or other groups, or 

 even in the plants of different countries. No character, however 

 important on some occasions, should be allowed to override all 

 others on all occasions. The valvate male calyx, for instance, to 

 which Mueller gives on most occasions so absolute a tribual value 

 as to make the most unnatural combinations, is never allowed 

 even generic value by Baillon, because of its inconstancy in 

 Crofon, whereas in many cases it certainly has no exceptions. 



After the publication of the ' Prodromus,' Baillon severely cri- 

 ticised some parts of Mueller's system in a paper in the eleventh 

 volume of c Adansonia,' in which he also described several addi- 

 tional genera and species, chiefly from the New -Caledonian col- 



• • * 



lections, then recently received at Paris. To these criticisms 

 Mueller replied with some bitterness in the ' Botanische Zeitung 

 for the year 1873, p. 229. He also, in a folio volume, which is 



