PARENTAL CARE AMONG FRESH-WATER FISHES. 405 



The most generally known English name for any species of Hoplo- 

 sterrtmii is Hassar, which, however, it shares with very different 

 species of the family of Silurids — the Doradines. The etymology 

 and origin of this name is unknown. In the great Xew English 

 Dictionary (Murray's) it is supposed to be of South American origin 

 (*' ? native South American name ""), but this is doubtless a mere sup- 

 position based on the fact that it is current in a South American 

 country (Demerara). According to John Hancock (1828) it is 

 derived from the "Arowaks." Another name having some currency 

 m English is one that may be often heard in the British island of 

 Trinidad. In that island the common species is generally known 

 as the Gascadura^ which is a survival from the time of the Spanish 

 dominance in that island. This has been translated as " hatdhead," 

 but it evidently refers to the segmented or articulated appearance 

 produced by the two rows of high and narrow plates which cover the 

 sides, the Spanish word Cascadura meaning " an act of l:»ursting or 

 breaking asunder." The same fish is also 

 called que que by the negroes of Demerara, in 

 imitation of the noise it makes when taken out 

 of the water, which, it is considered, " much 

 resembles the cry of a rat. 



The roundhead hassar {H Ojylostermtrii litto- 

 role) was supposed by Hancock to remain and 

 bury itself in the mud of the pool which was 

 drving up, and he especially stated that it was 



'. r \ . , . ,, , Fig. 41.— J [oplostenium pec- 



not ' known to attempt such excursions as the torale, showing armature 



flathead hassar resorted to, '' although it is caD- "*" breast. After Bouien- 

 able of living a long time out of its element."' 



A close relative of the roundhead, however, the cambota of southern 

 Brazil {GaUichthys afiper), was the subject of interesting observations 

 by an American resident in Bahia, Mr. Joseph Mawson. He had 

 six specimens of the fish and kept them mostly '' in a narrow-necked 

 tin of Mater, with some sand and mandioca meal at the bottom, for 

 five days," and they continued " active and vigorous," Mawson's 

 specimens were not full grown, measuring only 2 to 4 inches (5 to 10 

 centimeters) long. Occasionally he let them out of the receptacle, 

 and once one got out without assistance. Mawson's account was 

 published in Science for December 25, 1880 (I, p. 317), and is worthy 

 of partial republication here, especially as it has entirely escaped the 

 notice of all ichthyologists. 



I have had them out in the garden several times. I find that they move best 

 on smooth danii) ground, and are embarrassed by sticks or other inequalities. 

 They can jump a little vertically, but their progress on land is effected entirely 

 by a quick wriggling motion of the body, which is nearly flat upon the ground. 

 The paired tins (pectorals and ventrals) are extended laterally and seem to bear 



