A rectangular entrance 2 by 1 metres, open at the foot of a 2-metre drop, giving way to a single descending gallery totally occupied by clastic materials. This gallery reaches a width of 55 metres at some points, and the ceiling maintains a height of between 2 and 4 metres in the central part, while on both lateral sides of the gallery the ceiling lowers noticeably and some conduits can be found that may appear independent, but are no more than stretches of the main gallery separated from it by piled-up blocks. The whole cave develops very close to the surface, and some roots can be seen hanging from the ceiling. At the final end of the gallery there is a small descending conduit between blocks, at the bottom of which the maximum depth is reached. The current morphology suggests that its genesis is due to a flow of materials favoured by the dipping of the anticline in which it is located. The substrate on which the limestones rest is made up of marls and gypsum of the Keuper, and its plasticity facilitated the displacement of the limestones. Infiltrations caused a loss of volume sufficient to trigger a collapse process, giving rise to this singular chemoclastic phenomenon.

The absence of reconstructive forms (speleothems) is total.

Establishing a total length in a cave of these characteristics is always somewhat subjective. J. Borràs (1982) gives it 150 metres (probably following the discontinuity criterion); other authors have estimated the length at 300 and 350 metres. If we follow the data of the 1977 survey (from which we have the data taken in situ) and follow the continuity criterion, we have that, if we add the entrance-bottom polygonal line (by the usual and logical itinerary), plus the polygonal line of the lower galleries, plus the distances from the main polygonal line to access the small lateral galleries and the length of these, the total real length gives us a result of 265 metres.

Survey and speleometry: The two published surveys of the cave do not have a plan and are simplified to a general elevation and some cross-sections. Here we attach two unpublished surveys (SIE, 1971 and 1977) which do have the plan drawn. We update the record (December 2016) by adding a hitherto unpublished survey, also with the plan, made by members of the GE Pedraforca and SE Gelera in 1973 and which has been rediscovered after 43 years in the folder of old papers by Alfred Montserrat, who has provided it to us.