part 2] THE VALEXTIAIN'^ SEEIES. 159 



A few specimens of CJijnacograptus scalaris var. normalis have 

 been obtained from the Gasworks Beds. This variet}^ is stated in 

 the Pala3ontographical Society's ' Monograph of British Grapto- 

 lites ' to survive into the zone of Moiiograptus grecjarius (of 

 Lapworth), but never into the Upper Bii-khilL The Gasworks 

 Beds accordingly occupy a horizon not higher than the zone of 

 Moiiograjjiiis convolutus (of Marr & Nicholson, etc.). 



A single specimen of Glimacograptus was obtained in the strata 

 which are considerably above the base of the Bosemarket Stage. 

 The generic identification is certain, but its state of preservation 

 is too poor to allow of the species being determined. Despite this 

 fact, the discovery of any species of Climacograptiis is imjwrtant. 

 According to the ' Monograph of British Graptolites,' CI. scalaris 

 occurs rarely, associated with Has f rites maxim us and Monograptus 

 turriculatus, in the Lower Gala Beds, and the genus is not known 

 to occur anywhere above that horizon. It may be inferred that 

 the local base of the Bosemarket Stage descends to a horizon near 

 the base of the Gala or Upper Valentian Stage. 



Shelly fauna. — It has been customary to regard the well- 

 known fossiliferous mudstones of the Gasworks as the equivalents 

 of the Mulloch-Hill Sandstones, largely on account of the abun- 

 dance of the problematical Nidulites favus in both localities ; but 

 there are serious difficulties in accepting this correlation, and all 

 the evidence indicates that their faunal affinities are rather with 

 the Woodland Beds and the Newlands JPf utamerus Group. 



This conclusion is based mainl}^ on a comparison which I have 

 made between the brachiopod genera, Pentamerus, StricJcIand- 

 mia, Stropho)nena (sensu lato), and Plectamhonites, of the two 

 districts and the trilobite genus Phacops (sensu stricto). These are 

 represented by numerous forms, and a detailed study of the fossils 

 of the Haverfordwest district has shown that many of them 

 undergo progressive modifications in ascending order. They thus 

 acquire a greater value for the purposes of correlation than forms 

 for which no evolutionary history has been traced, however well 

 marked they may be. A brief summary of these comparisons is 

 all that can be attempted here. 



The Pentamerids are represented in both districts by JBarran- 

 (lella, Stricklandinia, and possibly Pentamerus. A variety of 

 ParrandelJa. undata which is found in the Mulloch-Hill and 

 Newlands Beds is characteristic of a horizon in the Cartlett Beds 

 of Haverfordwest. The form which is so abundant in the Wood- 

 land Beds is different, and is not represented in the southern 

 region (see, however, Llandovery district, p. 168). Pentamerus 

 ohlongus is recorded from the Newlands Beds ; but, as it does not 

 appear in Mrs. Gray's extensive collection, the record is open to 

 doubt. A single specimen of Pentamerus sp. was found in the 

 lower part of the Gasworks Mudstones. 



Smooth foi'ms of Strioldandinia which have commonh^ been 

 attributed to S. lens occur in the Mulloch-Hill Beds, and less 

 commonly in the Woodland and Newlands Beds. Similar forms 



