is that the Lingulae are found in abundance in the marine littoral fauna where they occur, normally associated with marine species of molluscs, crustacea, etc., in the marine deposits of the same age further south; furthermore, Lingulae are found from the Cambric to the present in undoubted marine associations. The eurypterids, on the other hand, are not found in the unequivocal marine deposits to the south, but appear quite as suddenly as the Lingulae, although in separate bands. They have been found to the north of the Ludlow area in Scotland, always as concomitants of the transition from, marine to continental conditions, and it is only when the latter conditions transgress farther and farther south that the eurypterids appear.
I think that much light will be thrown upon the interpretation of the late Siluric deposits in England by the study of the Ludlow and higher bone-beds. It will not be possible in this paper to consider the habitat of the early fishes except incidentally, but if that is proved to be fluviatile, as I think it may be, then the following explanation may be offered for the bone-beds. The Ludlow Bone-Bed, which is the most constant and widespread, appears to mark the wholesale destruction of the fishes in the rivers at the time when, in the oscillatory movements preceding and accompanying the retreat of the sea, there were temporary advances. The salt water, pushing its way up the rivers, killed the fishes and other river organisms in great numbers, for the fluviatile fishes can less easily survive an influx of salt water than marine fishes can an influx of fresh water. This is implied in Günther's statement that "On the whole, instances of marine fishes voluntarily entering brackish or fresh water are very numerous, whilst fresh-water fishes proper but rarely descend into salt water" (97, 187). Thus during the oscillations preceding a negative eustatic movement, the sea would occasionally advance a short distance over the land, and if this temporary positive movement were widespread, bone-beds would be formed at or near the mouths of many rivers almost contemporaneously, and even if some areas were submerged and others not, geologically the bone-beds would appear to be approximately synchronous. This theory is borne out by the occurrence of thin bone-beds at a number of higher levels in the beds above the Ludlow Bone-Bed. Moreover, whenever there was a slight retreat of the sea with the pushing forward of terrigenous, coarse material, then the light Lingula shells might well be left stranded along the line marking the high-water level for that particular period. If such a negative movement were followed by a
