August 20, 1903] 



NATURE 



379 



about the size of a small hazel nut. It consists of an outer 

 envelope and a central body, the two being closely joined 

 together, except towards the top, where the envelope leaves 

 a narrow passage open, leading down to the central body. 

 The ape.\ of the latter becomes excavated into a hollow 

 pit — the pollen chamber — a feature almost peculiar to 

 Cycads amongst living plants, discovered by our country- 

 man Griffith so long ago as 1854, though the credit is 

 often wrongly given to later French or German investi- 

 gators. 



The pollen, blown by the wind or possibly conveyed by 

 insects, is received in the opening of the envelope by a 

 drop of gummy substance, and as this evaporates the 

 pollen-grains are drawn down through the narrow passage 

 into the pollen chamber below. There each grain anchors 

 Itself by sending out a tube into the neighbouring tissue 

 of the ovule. Thus pollination is accomplished. Fertilisa- 

 tion, i.e. the actual union of the male and female cells, 

 takes place some months later, when the ovule, now to all 

 external appearance a seed, has reached its full size. In 

 the meantime, the single megaspore or embryo-sac, em- 

 bedded in the tissue of the central body of the seed, has 

 grown to enormous dimensions — filled itself with prothallus 

 and developed the egg-cells at its upper end, which are so 

 large as to be easily seen with the naked eye. 



The pollen-grain behaves like a cryptogamic microspore 

 and produces two large spermatozoids, each with a spiral 

 band bearing numerous cilia — the organs of motion. The 

 pollen-tube becomes distended with water, bursts, and sets 

 free the sluggishly moving spermatozoids, which by aid 

 of the water discharged from the pollen-tubes are able to 

 swim to the egg-cells and effect fertilisation. 



This remarkable process, first discovered in 1896 by two 

 Japanese botanists, Ikeno and Hirase, and independently 

 in 1897 by the American Webber, occurs not only in the 

 Cycads, but also in that strange plant the maiden-hair 

 tree. Ginkgo, a form now completely isolated, certainly 

 rare in a wild state, and said to have been only saved 

 from extinction by cultivation around Buddhist temples 

 in China and Japan, but which has a long geological 

 history. 



The cycadean method of fertilisation holds exactly the 

 middle place between the purely cryptogamic process, 

 where the active male cells accomplish the whole journev 

 to the egg by their own exertions, and the method typical 

 of seed-plants, where these cells are little more than mere 

 passengers carried along by the growth of the pollen-tube. 



The adaptations, which in the Cycads allow of pollination 

 and fertilisation on the plant, are chiefly three : — 



(i) The envelope of the seed with its narrow opening 

 down which the pollen-grains are guided. 



(2) The pollen-chamber below in which they are received. 



(3) The pollen-tube which, however, plays a somewhat 

 less important part here than in the higher flowering plants, 

 and in the Pahtozoic allies of the Cycads may perhaps have 

 been dispensed with altogether. 



There are, however, other points in which the ovule of 

 a Cycad differs from the spore-sac of a Cryptogam. Not 

 only is the megaspore solitary — that is a condition already 

 reached among the water-ferns — but it is firmly embedded 

 in the surrounding tissue. It is no longer a 'mere spore 

 destined to be shed, but remains throughout an integral 

 part of the ovule, while the ovule ripens into a seed and 

 ultimately germinates. Thus the whole development of 

 the prothallus takes place within the seed, and this requires 

 special methods of food-supply, involving a complexity of 

 structure far beyond that of any cryptogamic spore-sac. 

 When the time tor dispersal comes, the seed is shed as a 

 whole. 



There is, however, another character commonly regarded 

 as essential to the definition of a seed ; a seed should con- 

 tain an embryo. This implies that, after the egg-cell has 

 been fertilised, the young plant develops to a certain extent 

 while still within the seed, and before it is shed. In the 

 ripe seed the embryo passes into a resting stage, and only 

 resumes its development when the seed begins to germinate 

 and the embryo becomes a seedling. Usually, too, the 

 ripening of the seed itself is dependent on the development 

 of the embryo; if there is no fertilisation there is no true 

 seed, only an abortive ovule. 



In the Cycads this is not the case ; the ovule ripens into 

 a full-sized and apparently normal seed, even if fertilisation 



NO. 1764, VOL. 68] 



has failed. In our hot-houses Cycads are seldom fertilised ; 

 yet the conspicuous scarlet seeds of Cycas revoluta, or the 

 crimson seeds of Encephalartos, are familiar objects to many 

 Kew visitors. Further, the degree of development of the 

 embryo at the time the seed is shed is very inconstant ; 

 sometimes, although fertilisation has taken place, the 

 embryo is scarcely to be detected. 



The definite resting stage of the young plant in the dry 

 seed, so characteristic of the higher Phanerogams, is un- 

 known to these primitive seed-bearers, the Cycads and the 

 maidenhair-tree. The same appears to hold good for 

 the seeds found in the Pakeozoic rocks. Such seeds are 

 common in certain localities, as in the Coal-measures of 

 central France, and to a less degree in our own coal-beds. 

 In petrified specimens the structure is often beautifully pre- 

 served, yet in no single case has a Palaeozoic seed been 

 found to contain an embryo. It is not merely a matter of 

 preservation, for that is not unfrequently so good that the 

 delicate egg-cells can still be recognised. Thus there is 

 no known " seed " of Palseozoic age which, according to 

 current definitions, strictly deserves the name. Technically, 

 the term " ovule " would be more appropriate, but the 

 obvious maturity of the integument makes the word " seed " 

 seem more natural. So far the case is parallel to that of 

 our recent Cycads or the maidenhair-tree. 



It is, of course, possible that any day we may light on 

 some Palaeozoic seed with an embryo ; it may be that the 

 specimens hitherto found were all unfertilised, though the 

 frequent presence of pollen-grains in the pollen-chamber 

 makes this explanation unlikely. It seems not improbable 

 that the development of an embryo in the ripening seed was 

 a later device — that in the older seed-plants the period of 

 rest came immediately after fertilisation, and that the 

 growth of the embrjo, when once started, went on rapidly 

 and continuously to germination. In that case a seed with 

 a recognisable embryo would rarely be preserved. 



We are now in a position to see what are the chief 

 advantages gained by a plant in adopting the seed-habit ; 

 they are : — 



(i) Pollination on the parent plant, and consequently 

 greater certainty in bringing together the two kinds of 

 spore. 



(2) Fertilisation either on the plant or at least within 

 the sporangium, giving greater certainty of success, and 

 protection at a critical moment. 



(3) Protection of the young prothallus from external 

 dangers. 



(4) A secure water-supply during its growth. 



(5) Similar protective and nutritive advantages for the 

 young plant developed from the egg-cell. 



This last end, however, was very probably not yet fully 

 attained in the earlier seed-bearing plants. 



We may now go on to consider our main subject — the 

 historical question, from what group of spore-bearing plants 

 were the seed-plants derived? 



One thing is plain ; the stage of heterospory was the 

 immediate precursor of seed-formation, and it was from 

 some group of Cryptogams producing spores of two kinds 

 that the seed-plants sprang. Such heterosporous groups 

 are, however, known in three of the main phyla of the 

 higher Cryptogams. 



In the Lycopod series we have, among their living re- 

 presentatives, pronounced heterospory in Selaginella and 

 Isoetes ; among the Palaeozoic Lycopods it was commoner 

 still. Within the class of the ferns we have the hetero- 

 sporous water-ferns. In the third series, that of the horse- 

 tails, w^e have, it is true, only homosporous forms now 

 living, but in Palaeozoic times a well-marked differentiation 

 of micro- and megaspores was attained, though less extreme 

 than in the other two lines. 



So far, therefore, there is no reason why the early seed- 

 plants might not have had family relations with any of these 

 great cryptogamic classes, and, as a matter of fact, all three 

 lines have been championed by one botanist or another as 

 the probable ancestors of the seed-plants. 



The horsetail stock, though it attained an extraordinary 

 development, shows no further sign of transition towards 

 the higher plants. 



The case for the Lycopods is stronger, and, indeed, they 

 were long the " favourites," and were commonly regarded 

 as lying nearest the true line of spermophytic descent. 

 This idea was specially based on the mode of development 



