430 TRANS. OF THE ACAD. OF SCIENCE. 



tal placentae and one-celled capsule) or on a fold which 

 extends to the centre and forms the dissepiments (central 

 placentae and three-celled capsule) ; shorter dissepiments 

 make semi-trilocular capsules. Very fragile dissepiments, 

 which break off when the capsule opens, leaving the placentae 

 central and detached (septil'ragal dehiscence), are found in 

 J. repens (therefore generically distinguished by Desvaux), 

 and to some extent also in J. Parryi, J. patens, and J. 

 setaceus. The placentas of J. Ronnerianus are enormously 

 developed into a spongy mass, which tills the greater part 

 of the capsular cavity. 



The capsule opens almost always from top to the middle or 

 to the base ; only in some of our species with subulate capsules 

 (J. scirpoides, J. nodosus) the separation of the valves com- 

 mences in the middle, while at the top they remain united 

 for some time. 



Seeds. — The seeds, when perfectly ripe, furnish some of 

 the most interesting and constant characters, but they are so 

 small and their markings so delicate that only a strong glass, 

 or, better, a microscope with a magnifying power of fifty or 

 sixty diameters, will properly exhibit them. It may not be 

 useless to caution botanists not to be deceived by seeds 

 loosely lying about with the specimens, as they very often 

 will be found mixed. 



The seeds are ascending or (the elongated ones) more or 

 less erect, with a lower end at the insertion of the funiculus 

 and an upper one at the chalaza, both ends united by the 

 raphe and often by a distinct fold of the testa. The seeds 

 are usually obovate or oblanceolate, thicker at the upper than 

 at the lower end, mostly terete, or, in rare cases (J. trifidus), 

 angular, when a few large seeds are pressed upon one another. 

 The ends are sometimes obtuse (J. bufonius), but commonly 

 either abruptly or more gradually pointed, apiculate or even 

 fusiform {J. 2>allescens, nodosus, scirjwides). Very frequently 

 the testa is slightly elongated beyond both ends of the body 

 of the seed and forms a small, membranaceous appendage (J. 

 efusus, tenuis, marginatus) ; in such seeds the longitudinal 

 fold of the testa, mentioned above, also becomes more, dis- 

 tinct. In many, and apparently more in American and in 

 alpine or arctic, species {J. Drummondii, Greenii, Canaden- 

 sis, etc.) these appendages become more conspicuous, and 

 extend beyond the seed itself as empty, shrivelled, tail-like, 

 white, or, rarely, reddish sacs. Such seeds have been called 

 scobiform; their seed-coat is more loosely adhering and some- 

 times {J. stygius) can be readily removed. This elongation 

 of the testa is of great diagnostic value, but the absolute or 

 proportionate length of the appendages is extremely variable ; 

 even in the same capsule I find the lower seeds with shorter 

 tails than the upper ones, and in J. Ca?iadensis we see forms 



