their water tniin tlje iipjier courses of the Alniendares, thon.iili this is so 

 highly improljaljle tliat tlie siijii^estion may be left out of consideration. 

 The springs heing situated on the south side of the lowi-r course of the 

 Alinendares the region aci'oss the river— that is tlie region nortli of tlie 

 river — may be excluded as a possilde contril)uting source of tlie sui)i)ly of 

 the Vento Springs. The region about the springs is composed of corral- 

 line rock. In such porous material conditions under which territory on 

 one side of a river may contribute to springs located on the opposite side 

 of a river are impossible. 



The most probable origin of the Vento water supjdy can best be 

 understood after a general statement of the conditions of the sui'rounding 

 region. 



The southei-n slope of the provinces (Juanajai, Havana and Matanzas 

 is largely drained by underground streams. The streams arising in the 

 hills and incnuitains. forming the watershed Itetween north and south 

 drainage, run aliove gromid for a distance and then disappear under- 

 ground. The Arignanabo River thus runs into a bank at San Antonio de 

 los P.auos and disappears among fallen rocks. A few yards away from 

 its "'sumidero"' the water can l>e seen running in its underground channel 

 through an opening in the thin roof of the channel. A few yards fiu'ther 

 on a dry cave leads down to the water, which, at the end of the dry cave 

 disappears among fallen rocks. Other rivers disappear in a similar man- 

 ner. They can not be followed in their underground coiu'ses becaiise they 

 completely till them. The undc^-ground waters and the channels in which 

 they run can, however, be reached in places through sink-holes. The 

 streams reappear, in part, at least, in a numl)er of "ojos de agua," some 

 near the coast south of San Antonio. The region drained by underground 

 streams i.s comparatively flat with frequently no indications of surface 

 sti-eams and tlieii- erosion, and extends westward to near San Cristobal, 

 where the lirst permanent surface stream is observed. At Artimisa and 

 Candalaria stream beds contained pools of water nt the time of my visit. 



From San ("ristolial to I'inar del Rio there are many small perennial 

 streams. Eastward from San Cristobal the cave region has an unknown 

 extent. Poey limited it to the jurisdiction of (iuanajay, i)Ut it certainly 

 extends as far east as tlie meridian of ^Slatanzas, and fi-om reports prob- 

 ably beyond Cienfuegos. East of Rincon there are, however', frequent 

 river beds, all l»ut one of which were dry during the time of our visit. 

 This main cave region Ix'longing to tlie soutiiern slope sends a tongue 



