Oxford University Cave Club

Proceedings 10 : "Pozu del Xitu"

OUCC Proc 10 Contents

OUCC Proceedings Index

OUCC Home Page

Geology of the Ario Region

Cave Development : Dye tracing

by Martin Laverty and Kev Senior

When we visited the Cornion area in 1979 there was little time to consider the geology of the region we were caving in. The geological notes of OUCC Proc. 9 were based on observations made when tramping to or from the caves. The expeditions in 1980 and 1981 consisted of full time caving broken by periods of exhausted rest so a proper study of the geology has yet to be carried out.

These are just some further observations which hopefully will be useful as a background to the thoughts on cave development which follow.

The Picos de Cornion consist almost entirely of limestones which outcrop over a vertical range of some 2000m. This thickness is achieved by the repeated overfolding and thrusting of a less thick sequence.

The bulk of the Picos de Europa limestone is grey-white, massively bedded, and has very few fossils. The individual beds are generally impossible to distinguish because of the many faults and joints. The beds are certainly very thick although it is possible that many were originally a series of thinner beds which have slumped and been redeposited as thicker ones. Sedimentary structures to support this theory have not yet been seen. Coarse conglomerates (calci-rudites) occur occasionally within this unit but they are not of great extent.

The other main limestone is a generally darker, finely laminated Mountain Limestone. The laminae vary in colour from black right through to pure white. Small folds disrupt the laminae and these are probably slump structures formed when the sediment was still soft. These rocks are also seen in the Valle Extremero at a much lower altitude.

Laying uncomfortably on the limestones is a sequence of highly fractured shales with a few limestones. The limestones have been redeposited and show features characteristic of turbidities. See fig. 1.

Nothing much can be said about the general structure except that it is complicated. The turbidities show that the shale sequence is the right way up but much work would be necessary to make the same statement about the main limestone units.

The rocks at Ario in which Pozu del Xitu is developed are different from most of the surrounding limestones. The Ario limestones have been subjected to a mild metamorphism with some associated hydrothermal activity. Indeed small scale open-cast mining of copper ore (probably malachite) was carried on in the Jou de Ario in the early 60s. Only traces of this activity can be found today. The degree of metamorphism varies considerably over small distances and can be described as patchy. The limestones have been changed into a poor marble and there is evidence that some fault planes have been 'welded' by the heat and sealed in some parts. The fault which is clearly seen in the entrance depression of Xitu can be seen to disappear over 5-10 metres into a mass of quartz and calcite veins.

Another important effect of the metamorphism is that the limestones are crowded with perfect anthigenic quartz crystals. The crystals reach 10mm in length and what soil there is around Ario glistens with them. In the cave, quartz crystals stand out in veins and clusters around the walls so that equipment is quickly worn out. The sand in the Teresa Series is composed almost entirely of quartz crystals so clothing gets coated. The stitching in the seams of oversuits quickly wears out, and where material comes in contact with skin painful sores develop. I leave the location of the sores to your imagination.

The quartz occurs as fine 'stringers' which pervade the rock but also as discrete crystals within the limestone. These are not 'fed' by veins so are presumably formed from silica grains present in the limestone. This implies that the heat of the metamorphism was sufficient to melt the silica in the limestone and 'soften' the limestone to allow the quartz crystals to grow unimpeded within it. There must also have been water present as much of the quartz has been mobilised and forms the abundant veins. There are many phases of calcite veining which indicates that at depth the heat was sufficient to melt the limestone in the presence of water. Nearer the lakes, one of the fine Iceland Spar type veins of calcite has been partially analysed and found to contain over 1000 ppm of Fe (0.1%) and also of Mn - doubtless related to the manganese mineralization at Bufferera.

It is probable that the Ario 'marbles' represent a deeper level than the surrounding rocks so that some kind of graben structure exists.

On the old Trea path, just past the shepherds' huts, there is a curious spotted limestone of uncertain origin. It is unaffected by the metamorphism and appears to be faulted against the Ario marbles. The best outcrop is shown in fig. 2. The rock would seem to have been formed by some kind of sedimentary/tectonic process rather than an organic one. The rock lies close to the others which are more clearly limestone breccias and it acts as a reasonable marker horizon. It strikes 090 with a dip of between 40 and 70 degrees to the north.

Cave Development in Pozu del Xitu

by Martin Laverty and Kev Senior

Top : Dye tracing

The most important fact about Ario as far as cavers are concerned is that the Rio Cares flows 1300m lower, and between the Ario plateau and the gorge there is nothing but limestone! One of the first things that is apparent is the abundance of promising looking shafts and depressions. These are the remnants of some past large cave systems which are possibly waiting to be discovered. Most of the holes are choked with boulders and cavers searching in vain for a way down. Clearly, large shafts are prone to collapsing and being filled with debris but there are some well sheltered shafts (e.g. Pozu Tras la Jayada - 367m) which are blocked quite deep below the surface. Xitu has quite a large entrance but the narrow rift which leaves the entrance depression still has 2m or so above boulders. The rift takes you away from the surface shaft so the main entrance pitches are not so susceptible to infill from surface debris. This is why you can get underground in Xitu when everywhere else eventually chokes. Even in Xitu the entrance pitches choke eventually but we had another piece of geological good fortune which enabled us to carry on.

Another reason why Xitu has not been filled with debris is that it is on a col and there are only gentle concave slopes above it. Thus rock loosened from the mountains cannot roll as far as the cave.

As previously mentioned the entrance depression of Xitu is formed on a prominent fault which strikes about 075. This fault is 'sealed' by metamorphism to the SW but continues some way to the NE. At the base of the entrance depression the rift follows the fault to the head of the entrance pitches. The rift appears to be mainly vadose with a phreatic roof. It is a trench between 3 and 5m deep, deepening towards the pitches and having an inaccessible upstream section whence floods emerge.

The first five entrance pitches after Climax Rift descend the sloping lower wall of the fault plane. There are no signs of water action on the hanging wall which is a perfectly flat surface. The trenching and potholing on the lower wall is not extensive and can be explained by the relatively small amounts of water that flow down the entrance at present. The size of the cavity that the entrance pitches descend is larger than can be explained by present water flow. What then formed the initial cavity? The rift is not big enough to have carried the volume of water necessary to create the large entrance pitch cavity. We think that the entrance pitches are a modification of a tectonic cavity under flow conditions not too dissimilar to the present day.

After the last 9m pitch the route drops away from the entrance fault and descends other minor faults. Our route from the bottom of the 19m pitch is in fact the original passage and the blind pot is the first of many examples of stream capture in Xitu, although it is now largely blocked - yet John appears to have proved a connection by dropping his ropewalker and finding it again on the rubble filling the smaller depression by the 37m blind pot at the bottom of the entrance series. One of the pictures in Caving International No. 11 does, in fact, reveal an appropriate opening in the roof at this point. An interesting point is where the last entrance pitch cuts through an old phreatic passage. This is the stroke of luck which enabled us to get away from the entrance series and into the larger, more ancient passages.

The phreatic passage is developed on a gently dipping fault - fig. 3. There is about 0.5m of layered sediment in the passage. There is no 'V' notch cut so the passage must have been dry when the entrance pitch 'cut through' it. How did this actually occur? The vertical fault along which the last entrance pitch is developed must have been present when the phreatic passage was active yet there is no sign that water ever leaked into this fault. The passage must have carried water right across the vertical fault 'ignoring' any potential it had for stealing water. The inference must be that some renewed movement opened the fault sufficiently to capture drainage from the surface or, more likely, that a new surface water source opened it. This source was presumably glacial melt water, possibly concentrated at the col by the crevassing of the ice sheet cover at such a position of probable difluence - towards the Cares and towards Los Lagos. Possibly concentration within the calcite vein or to one side of it insulated the upstream phreatic tunnel from invasion at any stage, but overflow did take place down its continuation to Customs Hall, which is trenched to more than a metre deep at its upper end. However, it is not trenched below here, so maybe this marks the phreas level at the time of opening the entrance series. It is also intriguing to note that the blind pot does not seem to connect to the present cave streamway from its choke, even though its base is only about 10m from it. The survey also indicates that its choke could be below the level of the present stream at its nearest point, so maybe there is another set of passages waiting to capture the Traversity Streamway and fossilize the Trench/Sump series.

The phreatic passage continues into the larger, phreatic Customs Hall. This passage is some 4m wide and 2-3m high and must have carried a very large amount of water at one time. Strangely, when followed upstream, the passage disperses into a series of vadose passages. The downstream extension must be under a boulder collapse. A deep vadose canyon cuts down into Customs Hall from a phreatic passage some 10-20m above. This is the top of Ming Piece Passage and presumably continues in the roof upstream (Traversity Streamway).

Between Customs Hall and the active stream passage there are many abandoned vadose oxbows. The total amount of vadose down-cutting must be in the order of 60-70m. The present vadose conditions have therefore been operating for a considerable time or are unusually vigorous. This latter possibility is in fact very likely due to the quartz sediment derived from the limestone insoluble residue.

There seems to have been more than one phreatic passage in the early development of the cave. The main one is the one seen in Customs Hall, and perhaps in Snowcastle, with others in the active streamway, Teresa Series and William's Bit. There are other smaller phreatic passages at higher levels but these are probably just loops from the main drain.

From the survey the cave clearly follows a trend on 100 degrees then changes its trend. What causes the cave to follow this course is not certain. It certainly cuts across many prominent fault/joint planes which might have caused a change of trend. The general trend of the cave must have been 'decided' by the phreatic passages with the vadose development modifying this trend. In detail, sections of vadose passage follow the line of least resistance among the many fractures. The old phreatic passages do not seem to be influenced by these planes of weakness but follow some other influence. You can only see what this is from the nearest mountain. Looking down from Jultayu an indistinct pale band follows the trend of the cave. It is difficult to see exactly what this 'pale band' is. It is not a bed of paler limestone. Rather it seems to be an ill-defined fault zone now re-cemented by calcite and quartz. The fault zone is cut by another fault at the start of the CBW Series where the trend changes. This new fault continues towards the Cares gorge and is a very clear feature from Jultayu. The new fault is probably more recent than most and offers an easy route for water to the gorge. It is really a fault zone bounded by two major faults and filled with a well cemented fault breccia. Passage enlargement by solution in a phreatic system would have occurred at a greater rate on these faults because none of the others provide such an easy, continuous course to the local base level.

The phreatic passage must have carried a large amount of water, the source presumably being melting ice from various glaciations. Any remaining morainic material at Ario should be studied to test this theory. The vadose passages show at least three stages of rejuvenation which would usually be linked to the various glaciations so the phreatic passage must have carried water from these earlier thaws.

It is worth considering what a phreatic passage is doing 1100m above the present base level. The Cares Gorge appears to have been cut by the Rio Cares as the mountains were raising with some later modifications by ice. The phreatic passage is either about as old as the gorge or is the lower part of a big phreatic loop whose higher levels could have been eroded by later glaciations. In a vertical plane the phreatic passage rises and falls over a range of some 50-60m (fig. 4). The lower parts of loops would have been sumps for some time before being drained by continuous down-cutting. About 20m downstream from the point where you drop into the active streamway a climb into the phreatic tube leads to a prominent line marking the position of a standing water level. This can be seen in other places and implies that water flow ceased for some time, maybe during a very cold phase with no water input and an ice-blocked resurgence. The very definite 'level' of development achieved by Teresa Series and the Traversity and active streamways would certainly appear to be significant in itself, especially as there seem to be other cave fragments exposed in the cliff faces around the Canal de Trea at a roughly comparable level. It seems not unlikely that this could be related to a pre-glacial landscape before the cutting of the gorge.

Afterwards, vadose development gradually took over at the highest points of the phreatic passages and working its way down. This is very well seen at the bottom of the climb up to Snowcastle (fig. 5). The vadose activity was accompanied by a series of captures, the latest of which is to be seen just before Flat Iron where the present small stream disappears down a fist-sized plughole, presumably to emerge from the mud and rubble at the shaft's base. The great amount of vadose down-cutting evident in the lower half of the cave and in the numerous captures could well be related to the abrasive action of the insoluble quartz residue of the limestone - an additional erosive element of what one would imagine to be considerable import which does not seem to be available elsewhere. (For instance, the inability of vadose action to compete effectively with phreatic solution in the opening of new passages has been remarked on by Pete Smart in his study of the Andara massif where LUSS have worked so successfully). It is thanks to these that the Teresa Series has been drained; the survey indicates that if the active streamway did not flow down the Trench Pitches the Teresa Series would be sumped at its lowest point. The first capture occurred at the Bold Step and water drained into Chopper Pitch. Later there was another capture at the Trench Pitch route and this eventually joined at the first capture at Enterprise Series. The immaturity of this passage is demonstrated by the sharp, narrow trenches and the minor extra down-cutting of the upstream trench compared with that in the Teresa Series.

Once the phreatic passages drained stalagmite growth was possible. Most of those in the roof of the active stream passage are 'dead' whereas the upstream section has some active stalagmite along with some amazing moonmilk curtains which are as soft as cream cheese. Moonmilk also occurs on the walls of parts of the entrance series, chiefly the section beyond the first blind pot (below 19m pitch) capture. Association of moonmilk with alpine karst is well known. What appears to be hardened, laminated moonmilk is also found as remnants of a much more complete fill throughout New Orleans passage. This is somewhat similar to the moonmilk floor deposits in West Kingsdale's Valley Entrance Milky Way, but was several metres in thickness before most of it was reeroded. It appears to show current bedding in places also, unexpectedly showing flow back towards Teresa Series. Perhaps William's Bit acted as a sink for this water? Unfortunately no indications of flow direction have been reported from there. The best stal development is in the Snowcastle. The passage leading to Snowcastle is phreatic in origin and seems independent of the Teresa Series. Snowcastle itself is in a fault chamber and the block on which the Snowcastle formation has developed has fallen from the roof.

Most of the stalagmite below Flat Iron is long dead and some has been cut right through by the present stream to reveal over ½m of finely laminated flowstone in one place; however that a little further downstream is 'alive' and very fine. The absence of any significant vegetation cover means that one would only expect extensive stalagmite formation at depth.

There are remains of at least three phases of sedimentation in the active stream passage and these are reflected in the Teresa Series. The deposits consist of well rounded limestone pebbles in a matrix of quartz sand. They are extremely well cemented and form thick, imbricated ledges from the Cover Picture Aven well into Teresa Series, after which laminated sands are predominant, the significance of which was totally missed in 1979; fortunately enthusiasm took us back anyway!

The phreatic cave is lost where the cave changes trend (below Servicio) and it is not clear where it goes. Possibly it continues on the old trend and it would be worth looking for. Certainly from the point where the trend changes to the bottom of the cave vadose development is entirely dominant. Water falls from an aven upstream on the new trend. There is definitely more passage to find, possibly an extensive cave system leading to a higher entrance? (or, more likely, to 2/5 and 3/5.) It will be quite a climb however. The water that enters after here (the water from the large aven just below Servicio flows back towards Teresa Series; see Xitu: the cave) is now followed to the bottom of the cave though it swells and wanes in amount as small cascades join and small sinks rob water. At the Gap a massive collapse has occurred between two faults (filling an old phreatic passage?) and in the roof of the chambers below you can just divine the outlines of a phreatic passage. The depth of the vadose down-cutting is so great that access to any roof tube would require a great climb. Once below the big pitch we have to admit that we didn't feel too much like admiring the cave development and just concerned ourselves with handling the caving. The general impression is that the pitch chambers are far too big. The passages linking the pitches do not seem big enough to have carried the amounts of water necessary to have formed the pitches. Possibly the pitch chambers are old shafts which developed on the same line of weakness that the present stream passage follows.

There are some very sharp narrow trenches and a series of steeply sloping rifts which emphasise the increasing immaturity of the passage. There are also some extremely attractive vadose trenches cut through calcite-cemented breccias and the walls are often covered with helictites. One can't help thinking that somewhere in the roof is a big, old phreatic passage leading into the upper reaches of an Agua type resurgence cave. We can only dream about this but maybe someone will go back to look for it one day. The cave eventually sumps some 200m above the bottom of the Cares Gorge and I think we are unlucky to have been stopped so soon. There is a good chance that the present sump is quite short and represents a change of trend for the cave. It probably emerges only some 100m away in another cascade passage dropping the ca. 50m to the upstream sump of Grotte de Culiembro, whence the water flows to another sump to drop the last 100m down a small passage much like the turbine pipes of an HEP station; obviously this is an immature passage formed since the last major incision of the Cares from the level of Culiembro. The good stal to be found in caves at this level could well date this event. That there was a stable valley floor at about this (path) level is shown further towards Cain where the old stream sediments are preserved in the wall of the gorge.

To conclude, Xitu has shown that big deep systems are present in this region despite past scepticism. The trouble is that finding the right entrance is difficult. All the most promising shafts, some around 250m deep, seem to choke and you are lucky to break into a horizontal stream passage. Steady, careful inspection of all possibilities would yield results. A good way to start might be to climb a mountain and look for long, persistent faults which trend towards the Cares Gorge. Then look at every entrance on the fault. Concentrate on the ones farthest from slopes which could supply debris to fill the entrance. Good hunting, and remember that the only certain law of cave location is that 'Caves are where you find them'! It took OUCC 18 years to find the right hole - and we were lucky.

Pozu del Xitu - 1981 dye tests

by Dick Willis

Top : Cave development

The 1980 Expedition produced an inconclusive dye trace from the small stream at the head of Flat Iron Shaft to the resurgence in the Canal de Trea, below Ario. Since Xitu was certainly going to go much deeper than the altitude of this resurgence it was obvious that additional work would be necessary to establish the major resurgence for the cave.

The first group to arrive at Ario in 1981 walked up the Cares Gorge placing charcoal detectors in all the known resurgences, including Trea, in order to obtain measurements of the background fluorescence in the waters. These detectors were changed before the introduction of the dye into the cave and were subsequently changed three more times at approximately weekly intervals. As a further control detectors were placed up and downstream of the likely resurgence area in both the hydroelectric canal and the river in case the cave resurged directly into the canal in one of its many tunnels or into the river at a previously unknown site.

0.5kg of Rhodamine B was placed in the terminal sump on 5/8/81 followed by 0.5kg of Fluoroscein into the pool at the head of the Flat Iron (the site tested in 1980) on 7/8/81. No visual positives were obtained but later analysis of the detectors shows that Rhodamine had appeared at the Culiembro resurgence by 9/8/81 and that Fluoroscein appeared at this site (and in the river downstream at Camameña) between 16/8/81 and 30/8/81. All other resurgences were negative for both dyes.

Analysis of the detectors was carried out by Pete Smart and Hans Friedrich at the University of Bristol Geography Department. Following Smart and Laidlaw (1977) the values obtained for each dye were plotted onto a scattergram (Fig.). Positive results were indicated by any points more than two standard deviations from the line of best fit drawn through the known negative (I.e. background) values.


Thanks to Cathay Pacific Mulu '80 for the dye, OUCC for the invitation and hospitality, Pete and Hans for the analysis and Anni for the statistics.


Oxford Expedition to the Picos de Europa in 1979 & 1980, by D. Rose
Smart P. L. & Laidlaw I. M. S. An Evaluation of Some Fluorescent Dyes for Water Tracing. Water Resources Research 13 pp. 15-33, 1977.

Top : Cave development : Dye tracing