500 THE POPULAR SCIENCE MONTHLY. 



"VYe have not found this method in Boerhaave (1727), nor in 

 Lemeiy (1675), nor does Faraday, in his " Chemical Manipulations," 

 make any allusion to it. As a process of purification of solutions, it 

 seems to have been lost sight of in modern laboratory practice. 



We have made trial of the method rendered noteworthy by more 

 than two thousand years' practice, and find that, while the process is 

 quite slow, it has certain advantages. Chief among these is the fact 

 that, when the capillary siphon is adjusted, it requires no further at- 

 tention ; there is no pouring into a constantly emptying funnel, and 

 there is no possibility of overflow. When properly arranged, the last 

 drop of liquid passes from the upper vessel to the lower, and, except 

 in certain cases of extreme divisibility, the filtration is perfect. The 

 process seems particularly adapted to the purification of weak solu- 

 tions, as of mineral waters, where the insoluble portion is not to be 

 conserved. There is obviously much choice in the material of which 

 the capillary siphon is made. We have tried cotton, wool, lamp-wick, 

 and asbestus, and we find stout, silky wads of the latter most service- 

 able ; it is also useful for filtering very acid and alkaline solutions. 



The rapidity of filtration does not seem to be hastened by length- 

 ening the longer arm of the siphon, and is chiefly dependent on the 

 number of threads in the filter, and on their fineness. In one experi- 

 ment, thirty-two strands of cotton yarn filtered twice as fast as sixteen 

 strands, and only half as fast as sixty-four strands. Oil filters much 

 slower than saline solutions, and the latter much slower than pure water. 



We have used the expression "capillary siphoning" in describing 

 anethisis, and perhaps it needs justification. The ascension of the 

 liquid is due to capillarity, and the descent through the longer arm of 

 the siphon is in obedience to gravity. Siphoning is dependent on 

 atmospheric pressure, and can not be strictly applied to the present 

 case ; yet we find by experiment that, if the lower arm of the Avoolen 

 threads be raised to the level of the liquid to be filtered, the action 

 ceases, and the form of a siphon is at all events essential to the process. 



The quotations from ancient authors show that they must have 

 been quite familiar with capillary attraction, yet the first observance 

 of this phenomenon is attributed by some authorities to Franciscus 

 Aggiunti, physician to the Grand Duke of Tuscany, about the begin- 

 ning of the seventeenth century.* (Aggixmti died 1635.) 



In conclusion, the object of this paper is not to propose a return to 

 this ancient method of filtration, the modern rapid processes being 

 more in accordance with present needs ; but we have thought it not 

 altogether useless to call attention to a neglected process which can in 

 certain cases be employed with advantage. Where the object of filtra- 

 tion is to collect the insoluble portion, it is obviously of no value ; 

 whether the process could be advantageously used on a large scale 

 remains to be tested, 



* Gehler's " Physikalisches Worterbuch," article " Capillaritat." 



