24. DR. J. D. HOOKER ON THE STRUCTURE AND 



onlv one cell to an orarium containing one or two pendulous ovules. This tendency to 

 suppression of the ovaria, comhined with the constant presence of then* styles, and the 

 stylos of the suppressed ovaries being in all respects similar to that of the developed ovary, 

 and equally perfect, is a very peciiliar character, frequent in the Haloragece, though 

 not absolutely confined to them : it is very conspicuous in Gunnera. The greater tendency 

 to imperfection in tlie female than in the male flowers of Haloragece, is also a marked 

 feature shared by BalanophorecB. 



3. Between Gunnera and Lophopliytum the affinity is so close that the female flowers 

 of these genera might be mistaken for one another ; and the male flowers of Lophopliytum 

 in their two stamens, linear anthers and basal short filaments, are absolutely identical 

 with those of several species of Gunnera : in the subgenus Misandra especially, the male 

 flower often consists of two small sessile calyx-lobes, with two alternating stamens. 



4. If the female inflorescence of Gunnera and Lophophytmn be compared, the affinity 

 may be very easily pursued : in each, short conical branches of the flower-head project 

 laterally from a stout axis, and are subtended by a large bract, and studded with a dense 

 mass of flowers, which consist of an adherent perianth, no trace of rudimentary stamens, 

 two styles, and a one-celled ovary, with a pendulous ovule, whose integument, in ripening, 

 contracts an adhesion to the inner wall of tke cavity. 



5 The tendency to a dimerous or tetramerous arrangement of the parts of the flower, 

 so cojispicuous in all Haloragece, and in Gunnera, is common to the Kelosidece and Lopho- 

 phytecB. 



Griffith has suggested an affinity between Mystropetalon and Loramthacece, founded on 

 the form of the male perianth, and the opposition of the stamens to its lobes ; but this is 

 not borne out by the female flowers, which must be considered of the highest importance 

 in establishing affinities. Griffith further was ignorant of the true structure of the seed 

 of Mystropetalon, and supposed that the genus had no relationship with Balanophorece. 

 After much consideration, however, I have included that genus in this Order, for reasons 

 appended to some notes upon its structure ; and in which view I am likewise following 

 that taken by Mr. Brown (Linn. Soc. Trans, xix. p. 233, in note). 



I have not dwelt upon the character afforded by the extreme dissimilarity of the sexes 

 of Balanophorece, and which is also conspicuous in Gunnera and Halorctgece ; because it 

 is common to many other Orders, and indeed is perhaps a very constant accompaniment of 

 reduction in structure, or of a normally imperfect development of the floral whorls. 



Amongst the objections that may be urged against associating JBalanophorecB with the 

 epigynous Calyciflorce, the strongest will probably be considered to be derived from the 

 habit, and the imperfection of the foliar organs : with regard to the former, it appears 

 wholly valueless, as will be proved by a cursory inspection of many Orders ; and of these 

 none are so conspicuous as Haloragece, which, for its extent, is one of the most polymor- 

 phous in the vegetable kingdom, and further, one consisting for the most part of reduced 

 forms of OnagrariecR. 



The extreme simplicity of the structure of the seed and ovule is another point of some 

 importance, and may be used as an argument against the alliance I have joroposed ; but 

 there is a manifest tendency to such imperfection in the epigynous Calyciflorce, especially 



