are those obtained from tlie ovary and fruit, but it is 

 on these last that I think the greatest reliance can 

 be placed in grouping the species. 



The characters of the order may be thus briefly 

 stated. ^ Calyx tubular, or, as in Orobanche and 

 Mgmetiay wanting, its place being supplied in the 

 former by the bracteoles, and in the latter by a 

 spathe. Corolla more or less irregular, tubular. 

 Stamens didynamous, usually incluse ; anthers either 

 perfect with 2 parallel poUeniferous cells, or imper- 

 fect, one of the cells being sterile or altogether 

 wanting. Ovary superior, 1-celled, or spuriously 2- 

 celled or, rarely, perfectly 2-cened; composed of 2 

 or 4 or more carpels, usually placed, when two, 

 anterior and posterior to the floral axis, with the 

 placentae right and left; ovules very numerous. 

 Fruit a capsule; dehiscence loculicidal. Seed 

 small, testa spongy, scrobiculate ; embryo minute, 

 at the base of fleshy albumen. Of these characters 

 those derived from the seed, taken in connexion 

 with the habit of the plants will be found most 

 constant. The placentation differs in the different 

 groups and has furnished the basis for the following 

 division of the order into sections or sub-orders. 

 In Orobanche, for example, it consists of from 4 to 6 

 prominent parietal lines, each covered with numer- 

 ous ovules. In PhelipcBa ramosa and Lathrcea squa- 



their admission into this list may lead to further, and 

 perhaps more accurate scrutiny so as to leave no 

 future doubt of their proper location. If really Oro- 

 banchs, they, especially the last, form through Striga 

 the transition to ScrofulariacecB^ but judging from Sir 

 W. Hooker's analysis of the two genera, the former 

 only, which has two placentae in each cell, is truly 

 referable to this order, while Anlaya passes into the 

 other. 



The first section, constituting, I presume, the type 

 of the order, unless I have misunderstood its struc- 

 ture differs widely from the rest in having several 

 carpels, all the others having only two. I am aware 

 that this view is at variance with the received 

 opinions of Botanists, and therefore forbear to urge 

 it beyond calling attention to what, to me, appears 

 to be the true structure of the genera I have referred to 

 it, and which, if confirmed by more extended ex- 

 amination, will, I imagine, lead to its segregation as 

 a distinct order. The second and third of my sec- 

 tions respectively represent in this group Bigno- 

 niacem and Gesnereacece. The fourth approaches 

 Scrofulariace(B, fi-om which it is kept distinct by 

 its imperfectly 2-celled ovaries and by the placen- 

 tiferous margins of each carpel remaining distinct, 

 in place of coalescing with its fellow into a single 

 axile placenta. Should further observation prove 

 maria (respectively the types of these genera),'the these sections, which I find of easy application, well 



carpels coalesce at the points of junction, but the 

 placentiferous margins remain free, and are after- 

 wards reflected to the right and left, but do not meet 

 in the centre ; hence the ovary is only one, or half 

 2-celled, In M^india the placenta consists of two the limits of genera, 

 intricately lamellated bodies, and in Oligopholis of 

 two solid, fleshy ones which, in both cases, nearly 



uniformity 



henceforth 



ters, in the absence of sub-division, causing a defi- 

 ciency of good distinctive marks by which to define 

 ' e limits of genera. 



The following synopsis of the genera known to 

 me, either from examination of specimens or good 



fill the whole cavity and are covered on all sides figures, may serve to explain my meaning, 

 with ovules, but in neither cohere, so as to form a 



central partition between the 2 carpels. And lastly, 

 in HyohancTie, the inflexed portions of the carpels do 

 partially meet in the axis, forming a spuriously 2- 

 celled ovary, but the placentiferous margins still 

 remain free and, being reflexed, form 2 loose pla- 

 centae in each cell. 



Availing myself of these variations of structure 

 and placentation of the ovary, I propose grouping 

 the order, so far as it is known to me, under the 

 following sub-orders or sections. 



L OROEANCHEEiE. Ovary of several (?) carpels, 

 1-celled: placentae parietal. Orobanche^ Cislanche^ 

 ConopltoliSy Anoplanthus. 



II. ^GijyETiEiE. Ovary of 2 carpels, 1-celled: 



placentae pgpetal large, fleshy, lamellate or solid. 



JEgineiia, Oligopholis. 



III. Lathr^e^. Ovary of 2 carpelsf 1-celled; centae 4, parietal. 

 carpels partially cohering m the parietes, ihe placen- 

 tiferous margins remaining free and spreading to the 

 right and left, forming two broad lamellar placentae. 

 Phillip(Ea^ Lathr(za^ Epiphagus. 



IV. Hyobanchee-e. Ovary imperfectly 2-celled, 

 that is, the inflexed carpels only partially meet in 

 the axis, but the placentiferous margins, remaining 

 free and being reflexed, form 2 loose placentse in 

 each cell. Hyobanche^ Camphellia, Christisoniaj 



Haveya^^ Aulaya.^ 



The last two genera are doubtfully added to this 

 list, Mr. Bentham having already referred them to 

 ScrofulariacecB^ but as the habit and, apparently, the 



I. Orobanche E^. 



1. Orobanche. Calyx wanting or rudimentary, 

 bracteoles dilated at the base, calyciform. Corolla 

 tubular, bilabiate, lips | lobed. Anthers glabrous, 

 placentae 4-6, parietal. 



2. Ctstanche. Calyx bracteolate, tubular, 5- 

 clefl. Corolla tubular, sub-equally 5-lobed. An- 

 thers woolly, placentae 4, parietal. 



3. CoNOPHOLis. Calyx bracteolate, sub-spathace- 

 ous. Corolla ringent, upper lip large, vaulted, xmder 

 short, 3-lobed. Anthers pubescent Placentas 4, 

 parietal. 



4. AiTOpi^ANTHUs. Calyx ebracteolate, 5-cleft. 

 Corolla tubular, arched or bent at the base, equally 

 5-toothed, or sub-bilabiate. Anthers glabrous. Pla- 



II. ^Bgij^etieje. 



5. iEaiNETiA. Calyx and bracts wanting, their 

 place supplied by a loose spath. Corolla tubular, 

 5-lobed. Anther cells divaricate, both fertile. Pla- 

 centae lamellate. 



6. Oligopholis. Calyx tubular, 5-toothed. Cor- 

 olla tubular, 5-lobed. Anthers 2-celled, one sterile, 

 subulate. Placentas 2, fleshy, not lamellate. 



III. Lathr^e^e. 



7. LathRjEa. Calyx ebracteolate, 4-cleft. Cor- 



seed are not in unison with the rest of that family oUa ringent, upper lip entire, under smaller, 3-lobed. 



( 3 ) 



