472 



DR. E. H. ilANKIN ON THE 



metres per Ksecond and if their velocities tend to increase as soon 

 as they begin to glide downwards, it is obvious that they must 

 possess some means of checking speed. Otlierwise they woidd be 

 unable to land without accident. In ' Animal Flight ' two methods 

 of checking speed, one used for high-speed flight, the other used 

 for low-speed flight, have been described. Two analogous methods, 

 for slow and fast speed respectively, employed by flying-fishes 

 have now to be described. 



5. Method of checking speed in high-speed flight. 



This method, which depends on changes of position of the hind 

 wings (pelvic fins), has only been followed by me in species whose 

 wings are coloured. It would be quite beyond my powers of 

 observation to see the adjustment in those species whose wings 

 are more or less transparent. Of the species that have yielded 

 most opportunities for these observations may be mentioned one 

 with black hind wings and another whose hind wings have a 

 chocolate-brown colour. The observations about to be described 

 are very difficult to make, and the following account is, to some 

 extent, compiled from glimpses of parts of the adjustment seen 

 on difi'erent occasions. 



The flying-fish starts its flight with the hind wings directed 

 outwards aiul backwards and extended in the horizontal plane. 

 At some distance, perhaps 50 metres, from the end of its flight, it 

 may be seen suddenly to lower its hind wings (pelvic fins) so that 

 they point nearly vertically downwards. On my recent voyage, 

 when using a binoculai- of magnification 10, it appeared to me 

 that, when the hind wings ai-e in the down position, their plane 

 is parallel to the direction of flight. In this position the leading- 

 edge only of the hind wing would offer direct resistance to speed 

 ahead. But on a voyage in 1914, when I was using an Aitchison 

 binocular of the very unusual magnification 25, on two occasions 

 the hind wings, when in the down position, appeared to me to be 

 slightly rotated round their long axis so that their hind margins 

 approftched each other. On one of these occasions, during part 

 of the flight, one hind wing was less visible than the other. 

 This was probably due to its having been seen end on. When 

 the fish changed its course so that it was tra.velling directly 

 away from me the two hind wings became equally visible. With 

 the hind wings thus disposed it is obvious that the air passing 

 between them must be, so to speak, entering a funnel, and therefore 

 much resistance would be offered to speed ahead. It is probable 

 that these two dispositions of the hind wings are used successively, 

 for, within about a. second of tlie end of the flight, the two hind 

 wings are suddenly rotated through nearly a, right angle I'ound 

 theii- long axes. Tlie result of this ma,n(ruvre is that the plane 

 of the hind wings ac(]uires a. position a,t right angles to the 

 direction of flight. In the first down position the hind wings are 

 not fully expanded. In the last position they are expanded to 



