308 M. Dulroclicl''s Ohficrvalrons on the Fontaine Rondc, 



wliicli is fifteen paces in length, and six or eiglit in breadth. 

 The water issues forth uninterruptedly from the deepest lying 

 part of the bottom, but from the highest part it ebbs once and 

 flows once every six minutes. This spring, therefore, is not 

 intermittent, but periodical. Springs of this description are, in 

 general, but rare occurrences, and the phenomena they exhibit 

 have always attracted the attention of the curious. 



Long ago Heron of Alexandria proposed a plausible explana- 

 tion of the intermittence of springs, in supposing that there were, 

 in the interior of the earth, reservoirs of water provided with 

 natural syphons. This explanation answers well for most cases ; 

 hence it has been adopted by natural philosophers. If the in- 

 termittence is of unequal continuance, or the swelling of variable 

 height, and if these inequalities are repeated regularly and pe- 

 riodically, we explain them by supposing that there are many 

 dissimilar reservoirs, and that each has its peculiar syphon. 

 All this is possible, and art can, by arrangements of this de- 

 scription, produce appearances resembling those in nature. But, 

 however appropriate this explanation may be, we must not for- 

 get that it is a mere hypothesis, and that nature may have other 

 means, besides those already mentioned, for producing the in- 

 termittence of springs. The careful study of the Fontaine 

 Ronde has afforded me a proof of this. This spring, as already 

 mentioned, rises during three minutes and falls for the same 

 time, so that its periods have a continuance of six minutes. But 

 I remarked, on frequent visits to the place, that the water did 

 not always diminish in equal quantity. Generally, the highest 

 part of the gravel bottom was entirely exposed by this ebbing ; 

 sometimes, however, the water did not fall so low as to uncover 

 the gravel. These anomalies did not exhibit any regularity in 

 their recui'rence ; and it was therefore difficult to unite them 

 with the periodical regularity which must be produced by one 

 or more syphons. If the intermittence of a spring is caused by a 

 syphon, the reservoir must necessarily be emptied, by means of 

 the syphon, in a shorter time than it would be filled again by the 

 afflux of the water. If the afflux is increased, the reservoir is 

 more speedily filled, and then the period of intermittence is 

 shortened, l)ut the flowing out from the syphon is lengthened. 

 Lastly, when the afflux brings as much water into the reservoir 



