Recent investigations have led to the remarkable discovery that many fern-like plants of the Carboniferous rocks, hitherto classed as Cryptogams, were in reality seed-bearers, and thus intermediate between Cryptogams and Cycads, the most primitive of existing seed-plants. They have accordingly been placed in a special group "Cycadofilices," or "Fern-Cycads," and regarded as transitional types, the view that they are the remains of a natural bridge connecting the Ferns with the Gymnosperms having received wide support,[330] and at first sight this conclusion would appear natural and obvious. But here, as in other cases, the difficulty is that the seeds which have been found are all fully developed; there are none in the intermediate stages between true spores and true seeds; we have the finished article, but no trace of seeds in the making; which upon any theory of evolution must have been exceedingly numerous. Hence Dr. Scott tells us:[331] "The important discoveries of the seeds of the Pteridosperms scarcely touch the question of descent, for these organs are of too advanced a type to throw light on the probable derivation of the group." In this instance, therefore, as in others, it remains true that in no case is any trace found of rudimentary character in the earliest fossil specimens of any class. It is undoubtedly a further puzzle that some of the Carboniferous cryptogams which did not bear real seeds, yet simulated them, a habit not easily explained on evolutionary principles. _D._ _The Course of Evolution._ The evidence of Professor Vines quoted in the text (pp. 202, 237) receives a remarkable confirmation from that of Dr. Smith Woodward, Keeper of Geology in the National Museum of Natural History. Speaking before the International Congress of Arts and Science, St. Louis, U.S.A., September 22nd, 1904, he thus touched upon the same question, which he illustrated especially from the history of fossil fishes, which he has made his special study.[332] "It must be confessed that repeated discoveries have now left faint hope that exact and gradual links will ever be forthcoming between most of the families and genera. The 'imperfection of the record,' of course, may still render some of the negative evidence untrustworthy; but even approximate links would be much commoner in collections than they actually are if the doctrine of gradual evolution were correct. Palæontology, indeed, is clearly in favour of the theory of discontinuous mutation, or advance by sudden changes, which has lately received so much support from the botanical experiments of H. de Vries.