] PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 239 



and it will be very interesting to compare these forms on some future 

 occasion with those described in the foregoing paragraphs. 



It is a well-known fact that the number of vertebrae in the spinal 

 column of the Anatidcc is by no means constaut. Even genera sup- 

 posed to be quite nearly related may differ in this particular, so that 

 careful records in this direction are very much needed, and when a 

 sufficient number have been taken to insure absolute accuracy such 

 data will be of service. 



In the subjoined table I have but little to offer, but it is the result of 

 a careful count in each case, and will go to show some of the differences 

 referred to and the method of comparison. 



Species. 



Number Ot Vertehr-p tint 



vertebrae in v ?£° r I 



Pfirviral re beilr hee 



Sf " bs "° 



Olor columbianus . .. 



Spatula clypeata 



Anas discors 



Glaucionetta islandica 



Now, iu the case of Spatula and Glaucionetta, in the specimens before 

 me, the thirty-eighth vertebra 1 , though free and really a caudal, lies 

 within the grasp of the hinder ends of the iliac bones, whereas in the 

 Teals this segment is found one vertebra's leugth behind them or entirely 

 without their grasp. It will be seen, however, that this does not affect 

 the total count,it remaining forty-four for the first-named genus and but 

 forty-three for the Garrot. I mention this because specimens may yet 

 be found where this thirty eighth vertebra has united with the pelvis, 

 as from the position it occupies it is perfectly possible for it to do in 

 the genera mentioned. 



The general characters of these segments as they are exhibited by 

 most Ducks are very well shown in Spatula. 



The atlas has its cup perforated by the odontoid process of the atlas 

 vertebra, and is characteristics in having the lateral canals — a feature, 

 so far as I am informed at present, that is common to the Anseres. 



An open carotid canal is provided for by the sixth to the twelfth ver- 

 tebras, after which a strong median hypapophysis takes its place, and 

 this becomes tricornuted in the sixteenth segment and first dorsal, while 

 in the eighteenth and nineteenth it is a long median plate. 



The fifth and sixth cervical usually has the best-marked neural spine, 

 which is there a long, though not high, median crest. The lateral ca- 

 nals iu the first half of the cervical region are long and tubular, while 

 the parapophyses are co-ossified for nearly their entire lengths with 

 their sides. Anatidw possess the " heteroccelous " type of articulation 

 among the centra of the spinal column. A strong hypapophysis is found 

 on the second and third cervical vertebras, to be much reduced iu the 



