The Tracing of Underground Waters. 51 



each tube, and on this were placed columns of fine sand, respectively 

 7, 14, and 27 cm. in length. These tubes were set vertically, 

 and flasks, containing a solution of one part of fluorescein in 200,000 

 of water, were arranged over them, in such a way that the solution 

 was delivered into the tubes about 7 cm. above the surface of the 

 sand ; the level and hydrostatic pressure were therefore constant 

 throughout the filtration, and approximately the same in each tube. 

 The following table shows the results : — 



Length of sand column 7 cm. 14 cm. 27 cm. 



Time taken from starting filtration 

 until first drop fell from the 



narrow tube 17 min. 35 min. 90 min. 



Volume of filtrate before florescein 



made its appearance evident... 9 cc. 17 cc. 36 cc. 

 Time taken for the solution to 

 pass through unchanged (esti- 

 mated approximately) 115 min. 180 min. 350 min. 



These results indicate that the amount of fluorescein removed 

 was proportional to the amount of sand with which the solution came 

 into contact, and taken in conjunction with the fact already estab- 

 lished, namely, that the removed fluorescein could be washed out of 

 the sand, suggests that the florescein is mechanicallv adsorbed, by 

 the sand or soil, and probably the adsorption proceeds until an 

 equilibrium is established between the concentration of the fluorescein in 

 the film adhering to the surface of the grains, and that in the bulk 

 solution. This view — that the phenomenon is a surface one — is sup- 

 ported by the fact that when coarse sand (presenting a smaller sur- 

 face) was used, a smaller amount of florescein was adsorbed. The 

 process would appear to be comparable wath the dyeing of w^ool with 

 substantive dyestuffs. 



In connexion with the use of fluorescein for detecting a water 

 contamination, it has been suggested that the method is of little use, 

 because it does not follow that bacteria can go where fluorescein can. 

 But this is an objection, which, if valid, applies with equal force to 

 all methods wherein the contaminating connection is sought by means 

 of any substance in solution (cf. P. F. Frankland's method with a 

 lithium salt). 



In so vital a matter as a water-supply therefore, I hold, that if 

 a connection can be established by means of fluorescein, between a 

 spot known to be contaminated and the source of the water-supply, 

 such suppiv should be regarded as dangerous, and liable at any time 

 to give rise to a water-borne epidemic. 



In view^ of the adsorbing action which soils and sands have been 

 shown to exert on fluorescein, it seemed desirable to ascertain if 

 bacteria would percolate through a column of soil in a similar manner 

 to fluorescein. 



A preliminary experiment with an organism of the B. frodigiosus 

 group, led to no satisfactory result, but more success was obtained 



