NA TURE 



[May 30, 1907 



<talked ovules, imbedded among the interseminal scales. 

 Surrounding tiiis central body is the hypogynous whorl of 

 stamens, (used below to form a tube, and expanding above 

 into the pinnate sporophylls, bearing very numerous com- 

 pound pollen-sacs or synangia, filled with pollen. The 

 whole is surrounded by an envelope of spirally arranged 

 bracts springing from the upper part of the peduncle. 

 The general arrangement of parts is manifestly just the 

 -.ame as in a typical angiospermous flower, with a central 

 pistil, hypogynous stamens, and a perianth. The resem- 

 blance is further emphasised by the fact, long known, that 

 the interseminal scales are confluent at their outer ends 

 to form a kind of pericarp or ovary-wall. When to these 

 general features we add the practically exalbuminous 

 character of the seed, with its highly organised dicoty- 

 ledonous embryo, the indications of affinity with the higher 

 llowering plants become extremely significant. The com- 

 parison was drawn by Dr. Wieland in 1901, immediately 

 on his discovery of the hermaphrodite flower. The angio- 

 sperm which he specially selected for comparison was the 

 tulip-tree, Liriodendron. The elongated 

 ^trobiloid fruit, with many carpels spirally 

 arranged in the receptacle, no doubt 

 suggests similarity, and, on general 

 grounds, we should naturally look for 

 analogies among the less specialised poly- 

 petalous dicotyledons, such as Mag- 

 noliacese, in some of which the leaves of 

 the perianth are spirally arranged. 

 .Analogies may also be found in our 

 familiar Ranunculaceje, such as Anemone, 

 or, still better, the globe-flower (TroUius), 

 with its numerous sepals, or, again, in the 

 water-lilies (Nymphpeaceje). In certain re- 

 spects, indeed, the Bennettitean flower was 

 in advance of these more primitive dicoty- 

 ledons, as seen in the arrangement of the 

 stainens, which have abandoned the spiral 

 phyllotaxis of the other organs to range 

 themselves in a definite whorl, while at 

 the same time their stalks are fused into 

 a tube, thus becoming " monadelphous," 

 as in the inallows of our own flora. 



The flower, with its great stamens, 

 10 cm. long in some species, must have 

 been a striliing object when it opened 

 (Figs. 2 and 3). As, of course, we can 

 know nothing of the coloration of the 

 perianth and other parts, we cannot tell 

 how brilliant its appearance may have 

 been ; the bright tints of the carpels and 

 ovules in some recent cycads, such as 

 species of Cycas and Encephalartos, sug- 

 gest the probability that the attractions of 

 colour were not wanting to the more 

 elaborate flowers of the older Cycado- 

 phyta ; the possibility of a relation to the 

 insect life of the period cannot be ignored. 

 It is not my intention to push further 

 the comparison of the Bennettitean fructi- 

 fications with the angiospermous flower ; 

 the deeply interesting questions which 

 must suggest themselves to the mind of every botanist, 

 as to how far these manifest analogies are lilcely to in- 

 dicate an immediate affinity, will be fully discussed else- 

 where by others. Enough has been said to show that the 

 remarkable organs discovered by Dr. Wieland fully merit 

 the name of " flower," in the same sense in which we 

 apply it, in everyday language, to the flowers of our 

 gardens and fields. 



As stress has been laid so far on the points of agree- 

 ment with the flower of the angiosperms, some reference 

 must now be made to characters which indicate relations 

 in other directions. The structure of the gyncecium renders 

 it probable, if not certain, that the BennettitejE were still 

 gymnospcrms as regards their mode of pollination, for the 

 openings between the scales of the pericarp leave the micro- 

 pyles of the seeds exposed. One must therefore suppose 

 that the pollen was received by the ovule directly, without 

 th» intervention of a stigina, so that functional angio- 



NO. T961, VOL. 76] 



spermy had not yet been attained. This is, no doubt, a 

 primitive condition, but it by no means excludes an affinity 

 with angiosperms. Just as in Lagenostoma, the seed of 

 the pteridosperm Lyginodendron, the beali of the nucellus 

 was still the receptive organ for the pollen, in spite of the 

 presence of an integument,' so, in the Bennettitean flower, 

 the micropyle of the seed was still the receptive organ in 

 spite of the presence of a pericarp. The integument in the 

 one case and the pericarp in the other might be termed a 

 " prophetic organ " in the only sense in which such organs 

 exist, i.e. an organ which has not yet assumed all the 

 functions to which it is destined. 



The stamens, while by their arrangement and position 

 they suggest those of a typical angiosperm, carry us back 

 by their structure and form to the sporophylls of a fern 

 (see Figs. 2 and 3), so that the characters of the 

 flower as a whole may almost be said to bridge the gulf 

 between cryptogams and the higher flowering plants. The 

 fern-like characters, however, have probably come to the 

 Bennettite.-E, not directly from true ferns, but through the 



— Cycaih'oidt-a ingein. Plan of bisexual flower consisting of a central ovuliferous cone, 

 a hypogynous whorl of compound stamens, united at the base, and a s;;ries of spirally 

 inserted enveloping bracts, all shown diaerammatically on about the same scale as 

 Fig. 2, and as if pressed out flat. From Wieland's '* .\merican Fossil Cycads." 



intermediate group of the Paljeozoic pteridosperms. The 

 fact that the pollen-grains are borne in compound pollen- 

 sacs, or synangia, like those of the Marattiaceje among 

 ferns, is one of great significance." It is iinpossible to 

 emphasise too strongly the extraordinary combination of 

 characters which the Bennettitean flower presents, uniting 

 in itself features characteristic of the angiosperms, the 

 gymnosperms, and the ferns, and suggesting that the 

 passage from the Filicineas to the higher flowering plants 

 may have been (comparatively speaking) a short cut. The 

 complexity of this earliest known type of a true flower 

 indicates the probability, as Dr. Wieland points out,' that 



1 See Oliver a 

 Lagcnpstotua Lo 



>tt, " On the Structun 

 Phil. Trans. Roy. S( 



of the Pala 



>zaic Seed 

 197 (1904), 



2 '1 he general question of the relation of the early seed-plants to fcrn- 

 is discussed in my article, " On the Present Position of Palaeozoic Botany," 

 Progressus Rci Botanica;, Heft i. 1906. 



3 " American Fossil Cycads," p. 143. 



