74 MB. S. S. BTTCKMAlSr ON [vol. lxxvi, 



realize its importance ; but work like that cannot be done all at once, 

 on account of many practical difficulties, and I am now finding just 

 the same fault with my own earlier records : they are not as detailed 

 as could be wished for later work. 1 However, with a map of the 

 country much can be recovered as to the position of specimens 

 by the method of faunal analysis : it is the object of this paper to 

 illustrate that result by the application of the method to a small 

 area. 



Some information in regard to the brachiopod problem may be 

 gathered from Dr. W. D. Lang's detailed researches on the Dorset 

 Coast, but this only applies to Cincta -. specimens of this genus 

 are found, apparently not commonly, and certainly not well pre- 

 served. In regard to the Rhynchonellids, the evidence is veiy 

 doubtful — only two crushed specimens: they look like the Rli. 

 thalia I rostellata series ; that is all that can be said. 



The comparison is given in Table V. 



regard to the Portlandian, where there are many species of which Wiirtemberg 

 shows nothing- — presumably stratal failure ; and the other as regards the 

 Inferior Oolite, where there is a fauna manjr times richer than the German 

 works show, especially in the strata of the hemeras bradfordensis to zigzag. 

 In the strata of the stretch of country from Crewkerne to Milborne Wick, 

 about 16 miles, is an ammonite fauna of almost incredible richness and 

 variety. In the plateau which runs from Bradford Abbas about 2 miles east- 

 ward there is entombed in the concavum-discites zone a fauna not only of 

 ammonites but of other invertebrates, probably without a parallel for rich- 

 ness in the world. And of this area only a few cubic yards have been 

 excavated. The wonderful series of Sonniuia (Euhoploceras) illustrated in 

 my Mon. Inf. Ool. Amm. pis. 1, lvii-ciii, from this area are unmatched, if not 

 almost entirely absent from the Continent. This is the more remarkable, 

 because in the same bed with them are Hammatocerata which occur also at 

 Capo San Vigilio (Lago di Garda), Italy, where, however, is no trace at all of 

 Euhoploceraa. The interesting question whether this is due to stratal failure 

 or to dispersal failure musb be left for consideration in a later paper of 

 this series. 



Wiirtemberg also lacks the large Perisphinctids, Aspidoceratids, and Cado- 

 ceratids of the Corallian (Argovian), of which there is a remarkable series in 

 the region immediately north-east of Wessex, if not in Wessex itself , rivalling, 

 if not surpassing, the grand series figured by De Biaz from Trept (V, 1). But, 

 so far as English palasontological illustration is concerned, such a series 

 might have no existence — it would be sought in vain. 



The agreements and differences in regard to Jurassic ammonite faunas 

 between regions like Wessex and Wiirtemberg will be an interesting study for 

 the future. At present, not only is publication failure a hindrance, but in 

 many cases too comprehensive generic names prevent the facts from being" 

 tabulated concisely. 



1 Collecting inch by inch in thin beds which are condensed epitomes of 

 thick deposits elsewhere is an ideal ; but, in dealing with remanie deposits. 

 there is a danger. In such deposits reversal of fauna may have taken place. 

 Thus, it is obvious that if Beds 3, 2, 1 in descending order have been heaved 

 up and are being denuded to form materials for the making of Bed 4, the first 

 specimens to be laid down would be those of Bed 3, and the whole order might 

 be reversed. There is a possibility of the armatus bed of Badstock (leckenbyi, 

 Baasayan 7), which contains the sweepings of many deposits, showing some- 

 thing of this. There is also a possibility of duplication of faunal horizons 

 occurring under some such circumstances. 



