A NGIOSPERMS. DICOTYLEDONS. 46 1 



Melampyrwn and Glolularia. The first cell of the endosperm fills the middle portion 

 of the embryo-sac in Veronica, the Labiatae, Nemophila, Pcdiadaris, Plan/ago, Cam- 

 panula andLoasa; it occupies the lower portion in Loranthus, Acanthus, ; Catalpa, 

 Hebenstreitia, Verbena and Vaccinium' In Nymphaea, Nuphar and Ceratophyllum 

 the upper end of the embryo-sac is separated off from the rest of the cavity by a 

 transverse wall soon after fertilisation, and the further formation of daughter-cells 

 (endosperm) takes place only in that upper part which contained the egg-apparatus ; 

 but here the formation of endosperm so far differs from that in the plants enumerated 

 above, that it commences with free cell-formation (Hofmeister). 



By far the larger number of true parasites and saprophytes belong to the group of 

 plants in which the endosperm is formed by division of the embryo-sac, with the 

 exception of Cuscuta which produces it by free cell-formation. 



Only slight traces of formation of endosperm are found in Tropaeolum and 

 Trapa, according to Hofmeister. 



The important points in the formation of the embryo in Dicotyledons have been 

 already described in the introduction to the Angiosperms ; it remains only to 

 mention here that the seed in parasites destitute of chlorophyll and in saprophytes 

 is ' ripe ' before the embryo has passed beyond the condition of a roundish body 

 of cellular tissue not yet differentiated externally, as in Monotropa, Pyrola, the 

 Balanophoreae, RafHesiaceae and Orobanche. There are some plants also which are 

 not parasitic, in which the embryo is still but little developed by the time the seed is 

 'ripe' and is detached from the plant, and in which it is matured only after that 

 time, as for example Erigenia bulbosa 1 . This recalls the circumstances described 

 above in the case of Gtngko. 



The amount of variation in the degree of development of which the vegetative organs 

 of Dicotyledons are capable may be illustrated here by the example of the Podoste- 

 maceae, a family of plants inhabiting tropical streams, the members of which often 

 simulate the habit of Hepaticae. Fine plates of this highly interesting group are to 

 be found in Tulasne's Monograph 2 ; Cario 3 and especially Warming 4 have recently 

 published researches into the history of development and the anatomy of some of 

 the species. Warming found no stomata in those which he examined ; the cells 

 of Tristicha and other species contain peculiar concretions of silica. The roots of 

 Tristicha have no root-cap ; the species examined by Warming had the root-cap 

 feebly developed and somewhat on one side ; the roots are attached to the substratum 

 by root-hairs and also by peculiar organs of adhesion which look like roots and 

 are sometimes branched, but have no root-cap, are exogenous in origin and consist 

 only of parenchyma ; perhaps, as Warming suggests, they are roots modified in a 

 peculiar manner. The leafy shoots are formed on the roots and in acropetal suc- 

 cession ; they are endogenous in origin and their structure is dorsiventral. The leaves 

 of Tristicha hypnoides are formed of one layer of cells and have no vascular bundle, 

 and even in the stem the spiral vessels are destroyed as soon as they are formed. 

 It is a remarkable fact, that the conditions here briefly mentioned are found also 

 in many parasites or saprophytes ; the endogenous origin of the leafy shoots in the 



1 Hegelmaier, Vgl. Unters. ii. Entw. dicotyler Keime, p. 144. Irmish gives a similar account 

 of the seeds of Ramincuhis Ficaria which are rarely matured. 



2 Tulasne, Monographia Podostemacearuni (Archives du Museum, VI, 1852). 

 8 Anatomische Unters. d. Tristicha hypnoides (Bot. Ztg. 1881). 



4 Familien Podostemaceae, forste Afhandling : Podostcmon Ceratophyllum, Mniopsis Wedel- 

 liana, Mniopsis Glazioniana. Vidensk. Selsk. Skr. 6. Rackke, with a summary in French, 1881. 



