S. E. Bishop— Kilauea in April, 1892. 207 



IV. For ordinary values the following empirical law is ap- 

 proximately true — Retinal persistence varies inversely as the 

 logarithm of the luminosity. 



V. The values of retinal persistence in dichroic eyes is very 

 different than in normal eyes. For instance, light impressions 

 of red last much longer on the retina of red-blind persons than 

 on the normal, yellow somewhat longer than normal and the 

 other colors about the same as normal. With green-blind 

 persons, green impressions persist much longer than normal, 

 red a little less than normal and the other colors the same as 

 normal. 



YI. The very marked departure from the normal values of 

 retinal persistence in dichroic eyes for the region of their 

 lacking color sensation, affords a precise and convenient method 

 of determining color-blindness. 



VII. Within the range of these experiments, it seems prob- 

 able to a high degree that age increases the duration of retinal 

 impressions to a nearly equal amount in all regions of the 

 spectrum. 



In conclusion I wish to express my obligation to Mr. E. 

 Gordon Merritt for his very valuable assistance in taking ob- 

 servations in these experiments. 



Physical Laboratory of Cornell University, June, 1892. 



Art. XXYII. — Kilauea in April, 1892 ; by Eev. Sereno E. 



Bishop. 



[Communication to J. D. Dana, dated Honolulu, April, 30, 1892.] 



On a visit, a little over a fortnight since to the volcano of 

 Kilauea, the condition of Halemaumau crater was found to be 

 greatly changed from that in 1887, as described in your 

 " Characteristics of Volcanoes," and which I had the satisfac- 

 tion of observing in your company. It also differs so greatly 

 from what has been hitherto put upon record, that I venture 

 to send you the particulars, with some illustrations for such use 

 as you may choose. 



At the collapse of the crater on March 5, 1891, precisely 

 five years after the previous collapse in 1886, a deep open pit 

 was left as before. The great mound* had entirely disappeared 

 into the volcanic depths. The lava speedily reappeared with- 

 out apparent obstruction, and has since then risen to within 

 about 300 feet of the upper rim of Halemaumau, where it 

 formed a floor of remarkable smoothness. Perhaps two months 



*The mound referred to is the " debris-cone " which occupied the interior of 

 Halemaumau, the pit iu the southwest part of Kilauea. 



