October i i, 1906] 



NA TURE 



597 



TllK S'linV OF FOSSIL FISHES.' 



T^lll'; clisi'overy of general principles in the sludy of fossils 

 -'■ is much hampered by the imperfection of the geo- 

 logical record. As every geologist is aware, we are de- 

 prndent for our knowledge of the life of past ages on a 

 few isolated episodes which have been locally preserved. 

 There is no continuous history of the life of long periods 

 in the rocks of any region that has hitherto been well 

 explored. Cessations in the deposit of sediment, the re- 

 currence of unfavourable conditions, and extensive migra- 

 tions, among other causes, have all contributed to this 

 result. .\n increasing acquaintance with scattered episodes 

 in the secular di-velopment of life, however, tends to reveal 

 its main outlines; and if we are unable to discover 

 ihi- actual facts we can at least arrive at an 

 approximation to them which serves all immediate 

 purposes. If we can determine the " fashion," so 

 111 speak, which prevailed during each successive 

 period in the geological history of a race of animals, 

 we are able to distinguish between those changes in 

 anatomical structure which led to stagnation or ex- 

 tinction, and those which were necessary for evolu- 

 tion to a higher plane. -Vn acquaintance with the 

 precise links between one grade and the next is not 

 of supreme importance. 



In the case of fossil fishes, some general principles 

 are already discoverable, and they may be treated as 

 an illustration of the results which palreontology is 

 now attaining. 



The earliest remains of fish-like animals satis- 

 factory enough for discussion are those from the 

 I'pper Silurian rocks, both of FJurope and North 

 America. They suggest that long before the 

 l.ilter part of the Silurian period fishes had 

 alreadv become a flourishing and varied race, but could 

 not be preserved among fossils because they had not 



armour plates which were symmetrically arranged like 

 those of Pterichthys. 



No link is known between the Ostracoderms and the 

 typical fishes which have a lower jaw and paired fins ; and 

 it is evident that the latter had already appeared in Silurian 

 times before they possessed a skeleton hard enough to be 

 preserved among fossils. The .Silurian and earliest 

 Devonian .Acanthodians (Kig. 2), however, cannot be far 

 from the beginning of these typical fishes, and they seem 

 to show how paired fins began. These very old Acantho- 

 dians are known because Ihey are completely covered by 

 small, hard skin-granules like those of the oldest fossilised 

 Ostracoderms. Not only did the armour begin here in the 

 same way as in the OslrarodiMins, bul there was also an 



^__^'' 



[i. \.—Tlutoiins stotuns, Traquair ; head seen from ab_ ._ 

 twisted to show dorsal tin and heterocercal tail mainly in sjd« 

 about one half nat. sue. — tjpper Silurian ; Lanarkshire. To illi 

 the most primitive skeleton of separate 

 Traquair.l 



tubercle: 



acquired a hard skeleton. The Upper Silurian fossils show 

 how this skeleton first began, and, if we may assume that 

 the order in which the different kinds of hard parts 

 successively predominate is the order in which they evolved, 

 it is easy to perceive how they gradually arose. Fortu- 

 nately all the phenomena can be traced in one compact 

 group of lowly fish-like animals, the Ostracodermi or Ostra- 

 cophori, which are so readily distinguished from the fishes 

 proper that there is no risk of confounding with them 

 members of any other line of descent. The hard skeletal 

 parts were confined exclusively to the skin, and in most of 

 the earliest members of the group Jhey were merely 

 scattered tubercles of limy matter like the shagreen of 

 modern sharks (Fig. i). The tubercles fused together into 

 armour plates in two different ways. Sometimes (as in 

 the Cephalaspida) a few regularly spaced tubercles grew 

 larger than the others, and each of these becaine a centre 

 of attraction round which the immediately surrounding 

 tubercles coalesced to form polygonal plates. These 

 coalesced again in accordance with the shape and inotions 

 nf the underlying soft parts. More rarely (as in the 

 .Asterolepidai) fusion of the tubercles occurred first along 

 the sensory canals, thus eventually producing overlapping 



G. 2.— Outlines of Acanthodian Fishes, illustrating their 

 elongation in shape and loss of "intermediate spines," as 

 traced upwards in geological formations. a, Climatius 

 Egerton ; Lnwer Old Red Sandstone, Forfarshire, b, ^fes, 

 mUchclli (Egerton) ; ibid. c. .-icaHtlwiits sidcatus, Agassii 

 Carboniferous, lidinburgh. d, Acanthoses graiiiis. Roemei 

 Permian, Bohemia. [Figs. b. c after Traquair, d after 

 /I., ana] fin ; d.^ dorsal fin ; /'. .7V. , pairs of spines between p 

 ("intermediate spines"); /..pair of pectoral fins; ;■., pair 

 fins. 



gradual 

 they are 

 sculiger, 

 tcanthus 

 '. Lower 



Fritsch.] 

 tired fins 

 3f pelvic 



1 .Abridged from the Presidential Addn 

 February 2, igo6. (Proc. Geol. .Assoc, v 



NO. 1928, VOL. 74] 



> the Geologists' Association, 

 i.x., pp. 266-282, figs. 1-15.) 



occasional fusion of the skin-granules into plates where 

 stiffness was possible or necessary. .\ iew rows of the 

 granules fused together at the front edge of the median 

 fins above and below the body, thus forming cut-waters or 

 "spines"; and as a double series of exactly similar 

 spines " occurs along the lower border of the abdomen 

 where the two pairs of fins are found in later fishes, it is 

 reasonable to infer that these are likewise the stiffened 

 front edges of fins. In other words, paired fins were not 

 originally restricted to two pairs, but formed a double 

 series along the entire length of the abdomen. The later 

 Acanthodians (Fig. 2, c, d) had only the ordinary two pairs 



