LUMINESCENCE IN THE SEA 215 



time to time, for the most part in late summer, August and 

 September. An unusually late and very extensive visitation 

 occurred in December, 1919, when in some parts of the 

 Barrow Channel there was a well-marked brick-red oily 

 zone on the beach caused by the stranded Noctiluca, and 

 a bucket of the shore-water was compared by observers to 

 " thick tomato soup," and after the sea-water was drained 

 off it was found to contain fuUy 2,000 cubic centimetres 

 of Noctiluca. Some of these placed in a smaU aquarium 

 retained their power of luminescence for three weeks. 



Noctiluca has been known as a common cause of lumines- 

 cence in coastal waters for at least two centuries. In the 

 middle of last century, A. de Quatrefages made notable 

 observations on Noctiluca, in which he showed that the 

 light was emitted from weU- defined patches or slowly moving 

 areas of the surface; each composed of a large number of 

 scintillating points. 



Many of the Radiolaria, both simple and compound, also 

 show bright luminescence. 



CcELENTERATA. — Many of the Hydroid Zoophytes, the 

 Medusse, and the Alcyonarian Corals show brilliant lumines- 

 cence. There is no need to mention aU recorded cases, or 

 even groups : a few examples will suffice. Some of the 

 Medusae are responsible for the large spots of light, as large 

 as a coco-nut or a tea-tray, sometimes seen by voyagers, 

 especially in warmer seas. Once when at anchor, in a 

 native boat, on the pearl banks of the Gulf of Manaar, in 

 an intensely dark night, I saw the black sea around us in 

 aU directions lit up by an innumerable assemblage of what 

 looked like globes of fire, waxing and waning in brightness, 

 all simultaneously glowing and then fading away into dark- 

 ness, and after a few seconds Hghting up once more. This 

 periodic display continued for about an hour and then 

 disappeared. Unfortunately, we were fixed to the spot and 

 had no small boat, so it was impossible to capture a sample, 

 but the impression produced was that the phenomenon 



