ILLUSTRATIONS OF INDIAN BOTANY. 41 



clas««, of sufficient magnitude too, to admit of its being further distrihutecl either circularly or 

 olh^rvyise according to the taste or abilities of the expounder. Unfortunately the author. 

 (I,indley) not seeing the true value of this bond of union, has, in his second edition, chaneed the 

 arrangement much for the worse. Von Martius places Chenopodiaceae, JVyctacrineae &c &c 

 in one poxx\>—Amar ant nceae, and Paronycheae,\n a second and Caryophylleae, Ind Port ul'aceae 

 in a third— Lmdley m his second Edition has Ficoideae in one groxi^—Portulaceae Si/enaceae' 



second — and at a great 



distance ^marantnceae, Chenopodiaceae, and several genera referred by all other authors to Por- 

 tulaceaSy or Ficoideae^ in a third. 



These examples are 1 presume sufficient to establish the intimate relationship existine 

 among all the plants possessing this structure, and the impossibility of ever devising a satisfac- 

 lory distribution of the genera ipto natural orders that does not set out from that point 



The three orders above named, being united by this structure and being the only ones an- 



m 



judicious course to group them together as a sub-class or large order and then redistribute the 

 Indian genera info suborders ; leaving for future consideration the final determination of all the 

 orders into which the group ought to be divided, and the station which it ought to occupy ia 

 the system of plants. To attempt more than this would be inconsistent with the plan of this 

 work, which does not profess to offer a new arrangement but merely to illustrate that in use 

 indicating from time to time, as opportunities offer, those points which appear to me defective 

 and how they may be amended. 



Portulaceae. 



This as now understood by both Meisner and Endlicher is a large and apparently nolv- 

 morphous order, including several tribes the affinities of which seem rather remote Both 

 tliese author^ place here, not only the plants referred by us to Ficoideae, but also the genus 

 MoUugo. This seems a questionable association, though for the present I adopt if, as the true 

 Portulaceae have a central basilar placenta (the seed attached to the bottom of the solitary cell 

 of the capsule): but in Mollago and Glinm (excluding G. tria7ithimoides) the capsule has 

 several cells with numerous seed, attached the whole length of the axis ; having besides, a five 

 not two sepaled calyx and being destitute of proper petals. This case supplies another exam- 

 ple of the.necessity of viewing the whole of those genera, having the peculiar curvembryose 

 seed of this tril,e, as belonging to one great natural family or class and re-distributing them 

 Without reference to what has been already done, according as they can be most naturally ^rounl 

 ed into sub-classes or alliances, duly bearing in mind, that the order or class is as it comes from 

 nature s hand a truly natural group and that the object in view is merely to distribute its con- 

 tenfs in the most easy and convenient order. Until this course is adopted, it seems most im- 

 probable we shall ever hare then disposed in such a series as will give general satisfaction. But 

 before this can be done, the whole must be carefully re-examined. In the mean while as I can 

 see no sufficient reason tor altogether rejecting the views of these authors, though I confess I 

 cannot fully agree with either, I shall modify the arrangement of our Prodronms by brinaing 

 !sesuvmm homh,co,deae, where it does not associate well with MesembryaiUhemnm, [o Porta- 

 laceae ; and G///;?v5, our only other genus of Ficoideae, being much less entitled to a place there 

 than even hemvr„m I also remove and place with Mollu^o in a distinct section, which may then 

 be given to Paronychnceae as better agreeing with it, both in habit and in characters, than with 

 ^ortvlaceae. In this way, the order Ficoideae is removed from the Indian Flora, and justly so 

 far as these two genera are concerned, as it is clear that ^esuvium cannot be separated from Tri. 

 anthema as an order, and nuth^r it nor Glinus can very well keep Mesemhryanthemum 

 pany. Let us attempt to distribute the Indian genera according to ihese view's. 



The Indian curvembryate genera referable to DeCandolle's class (Zaly 

 number and may be thus distributed. 



com- 



'/ 



PoRTULArKAK.-raIyx2 parted. Corolla 4-5 petaled exceeding the calyx, conspicuous 

 L.apsufe one-celled circumscissiie, or opening by valves : phcenta central in the base of the cell' 



seed attached by diiitmct podosperms.—Succulent herbaceous plants. Portulaca, Tatinum ' 



