18 SUPPLEMENT TO THE 



in these specimens of C. Julia there is snfficient contrast between the colour of the loop and 

 the surrounding matrix to make the different parts of the loop very clear when revealed 

 opaquely and by fracture. In most of the specimens the loop is of a rusty-brown colour, 

 surrounded by a bghter-coloured matrix. Mr. Glass's preparations plainly show that 

 the median })late of the loop in C. Jidia is of the exact shape described by Winchell. In 

 some cases, when this plate is shown by fracture, it seems to have been split laterally in 

 half, as if it were really, as Winchell describes, a double plate, but so closely appressed 

 as not to show its being double by a transverse fracture. Winchell says, in his note to 

 me, " The appearance is, that the loop abuts squarely against the flat surface of the plate 

 on each side. Homological considerations, however, would render it probable that the 

 loop is reflected abruptly and passes into the vertical plate, or rather forms the vertical 

 plate. This then should be double. But this is only conjecture." On either side of the 

 front border of the median plate (^. e. that which is connected with the loop, and inclined 

 towards the anterior margin) there are some faint ribs or plications, and these are indi- 

 cated in Winchell's profile figure^ In addition to this the whole of the median plate is 

 finely striated longitudinally on either side, the striations having a slightly waved 

 appearance. One of the specimens which Mr. Glass has fractured seems to show the 

 crura near the origin of the loop. Mr. Glass could not ascertain with certainty the existence 

 of the " obliquely conical pustules," referred to by Prof. Winchell as occurring on the 

 surface of the loop and median plate, but he thinks there are some indications of them 

 on the median plate, and that possibly they might be seen in larger specimens, or with a 

 higher magnifying power than he has used. Erom the results thus obtained I agree 

 with Mr. Glass that Prof. Winchell's description of the loop in Centronella is on the 

 whole correct, with the exception of the crura, which are more easily shown in our 

 Devonian specimens of C. virgo, and which are described both by BiUings and Rominger 

 as occurring in C. (jlans-fagea. In the preparations of C. virgo the loop can only be 

 seen by transparency, and the median plate can only be shown with any certainty as I 

 have figured it. Probably in some cases this plate was imperfectly developed, or had 

 been partly broken away before fossilisation. In some cases, however, as I have previously 

 stated, Mr. Glass found indications of this plate being in the shape figured by Prof. 

 Winchell ; and the preparations of C. Julia referred to above seem to show that this was 

 the normal condition. I have already given Winchell's figures of C. Julia, in order that 

 the above references may be more clearly understood. 



