Farr and Florance. — Radio-activity of Artexiaii Water. 189 



been made, for the same water trickles on to a further tank in which gold- 

 fish are kept quite successfully. 



Character of the Water. — As far as dissolved salts are concerned, the 

 artesian water of Christchurch is remarkably pure, being used for many 

 chemical purposes for which distilled water is usually necessary. Indeed, 

 it has been quoted as an example of what an ideal water should be ; but 

 its character in this respect has no bearing on the question, as the cause 

 of the effects under consideration rapidly escapes. In another paper* we 

 deal with the gases dissolved, and with the way these gases are altered by 

 rippling over obstacles. Messrs. Marsh and Goreham have (Report, Bureau 

 of Fisheries, 1904, AVashington, p. 345 et seq.) ascribed what appear to be 

 somewhat similar results to an excess of dissolved gas, particularly nitrogen, 

 and say that 2 cubic centimetres of gas per litre in excess of the saturation 

 amomit will cause symptoms, whilst a somewhat larger amount may be fatal. 

 At present, however, we are concerned with the evidence in favour of 

 radium emanation being a possible cause, as this was not considered by them. 



We therefore made experiments to ascertain how rapidly the radium 

 emanation escaped as the water fell over obstacles. Tank No. 1 has already 

 been stated to have a value 197. In the very short distance from No. 1 

 to No. 2 the radium-emanation number fell away to 171, whilst the tank 

 in which the goldfish live, and to which a very small trickle goes, gives a 

 leak per litre per minute of 30. It thus appears that the emanation escapes 

 rapidly ; but, to test it further, some water from another well giving a 

 value of 127 was poured three times from one glass to another, whereby 

 the value became reduced to 39. 



By the courtesy of the authorities of the Acclimatisation Society we 

 h.ave]had facihties for observing the behaviour of 75,000 healthy wild brown- 

 trout ova as they developed from the eyed state until they hatched, and 

 of ascertaining how the mortality amongst them varied with distance from 

 the well. For this purpose they were distributed in two parallel rows of 

 five boxes in a row, each box containing 7,500 eggs. The water trickled 

 down these two rows from box to box, falling from one box to the next 

 over a fall of about 8 in. By this means the water, after supplying the 

 eggs in one box, was re-aerated before it went into the next. It was known 

 that there was a much heavier loss of eggs in the boxes the nearer these 

 were to the Avell — so much so that the use of the top pair of boxes (one box 

 in each row) had been discontinued owing to the great loss in them in the 

 egg stage, and also to the greater development of " blue swelling " in the 

 yolk-sac stage. The number of eggs taken dead oiit of the various boxes 

 was counted for us by Mr. Rides whilst we made a determination of the 

 radium- emanation content per litre and also {loc. cit.) of the gas content. 

 In the following results the dead eggs in any corresponding pair of boxes 

 are added together, and they are therefore the deaths in 15,000 eggs : — 



First pair of boxes . . 



Second ,, 



Third 



Fourth 



Fifth 



* " On the Influence of Ripples on the Gas Content of the Artesian AVaters of Christ- 

 ohurch." (See p. 237 of this volume.) 



