Fishes of the Western North Atlantic 485 



Si%e. This family includes the largest of the batoids, some of them (genus Manta) 

 growing to a width of more than 20 feet and to a weight of more than 3,000 pounds. 



Development. The Devil Rays are ovoviviparous. The uterine wall of the gravid 

 female is covered with the vascular villi '^ which are characteristic of ovoviviparous Rays 

 in general and which discharge a milk-like fluid on which the embryos are nourished 

 during the later stages in their growth. 



Habits. To a large extent the Devil Rays appear to have abandoned life on the 

 bottom to swim mostly near or at the surface. Like the myliobatid Rays, they progress 

 through the water by a flapping motion of the wing-like pectorals which suggest flying 

 rather than swimming. They often leap above the surface or somersault in the air, fall- 

 ing back in the water with a loud splash; sometimes companies of them do so together 

 or in rapid succession. However, the feature that chiefly distinguishes the habits of the 

 mobulids from those of other myliobatoid Rays is that they subsist on small pelagic 

 organisms, chiefly smaller schooling fishes and planktonic Crustacea. These they gulp 

 into their capacious mouths, apparently employing the flexible cephalic fins "to assist 

 in scooping the food into the mouth,"^* as described in more detail on p. 493. Old 

 tales that they use these fins to grasp swimmers or objects such as the anchor lines of 

 boats appear to have a foundation in fact (p. 509). 



Range. Tropical to warm-temperate belts of all oceans, including the Mediter- 

 ranean, both in continental waters and around offlying island groups. 



Genera. By common consent, primary subdivision of the family is based on the 

 position of the mouth, across the front of the head (terminal) or on the lower surface. 

 On this basis the group has been subdivided recently into two families" or subfamilies.'* 

 We follow the older and more usual course of regarding this distinction as a basis for 

 generic rather than family grouping. A secondary generic character is whether there are 

 teeth in both jaws or in one only, upper or lower. On the basis of these criteria, two long- 

 known genera are generally recognized: Manta with a terminal mouth and with teeth in 

 the lower jaw only; Mobula with the mouth on the lower side and with teeth in both jaws. 

 A third genus, Ceratohat'ts Boulenger 1897, has been based on a single specimen with 

 the mouth on the lower surface as in Mobula but with teeth in the upper jaw only 

 (p. 500). The genus is accepted here provisionally, but the possibility must be recog- 

 nized that the specimen in question may have been an abnormal Mobula. A fourth genus 

 of large Devil Rays with teeth in both jaws has recently been named Indomanta Whitley 

 jg^g 69 Apparently its mouth is terminal, as is the case in Manta., but the original 

 illustration'" of the specimen on which it was based leaves this point in doubt. If its 

 mouth were on the lower surface it would appear to have been a large Mobula. 



65. See RadclifFe (Bull. U. S. Bur. Fish., 24, 1916: 280) for an account of the uterine wall. 



66. Norman and Fraser, Giant Fishes, 1937: 83. 



67. Ceratopteridae, with mouth terminal, and Mobulidae, with mouth inferior, by Whitley (Aust. Zool., S [3], 1936: 

 165), who regards the mobulids as a superfamily rather than as a family. 



68. Mantinae, with mouth terminal, and Mobulinae, with mouth inferior, by Fowler (Proc. Acad. nat. Sci. Philad., 

 55, 1934: 243; Bull. U.S. nat. Mus., 100 \i3\, 1941: 479)- 



69. Aust. Zool., 8 (3), 1936: 183. 



70. Tombazi, J. Bombay nat. Hist. Soc, Jj, 1934: 227, Dicerobalis eregoodoa Tombazi; apparently not D. eregoodoo 

 Cantor (J. Asiat. Soc. Beng., 18, 1849: 1420), which equals Mobula diabola Shaw 1804. 



