Le Piccole Terme, an Analysis

Academic research information about Le Piccole Terme – the buildings open to the world in front of the Great Antrum – is not easy to find. Maura Medri is a highly-respected Professor at Università degli Studi Roma Tre. She is the authoress of many papers and books on archaeology.

Professor Medri has written an undated paper called “In Baiano Sinu: il Vapor, Le Aquae e Le Piccole Terme di Baia”. It appears that Professor Medri is relying on the academic authority of those who came before her for the purpose of the tunnel complex, for she says (translations from Italian by this website):

“By way of precedent, Cairoli F. Giuliani had correctly recognised the thermal function of the small baths and also hypothesized that the tunnels served to channel the emanations of the natural springs.”

and also:

“The description of the operation is not entirely precise, but overall Nielsen understood the system perfectly. At the time of building it was anonymous and it was described by Paget by the fantastic name, “Great Antrum”, also his work was not quoted in the bibliography because it is considered unreliable.”

So Professor Medri takes on face value what others said before her and discounts Doc Paget. However, she also says…

“Thanks to the kindness of Paola Miniero, I met Graziano Ferrari and Raffaella Lamagna, passionate cavers… on the basis of a shared project, after many years a working group has been formed that hopes to deepen the knowledge of the typically Baian heating systems. For now, I anticipate that Miniero and Ferrari will have confirmed to me that the sketches published by Paget are not to scale but plausible overall, although not entirely correct in detail.”

Did Giuliani ever set foot in the tunnels? Did Nielsen ever set foot in the tunnels? It would seem doubtful, for had they done so, their opinions might have been different, unless some serious cognitive dissonance is at play.


In the intervening years since Paola Miniero allowed further access to the tunnels for a new exploration and survey in 2014, very little has happened.

The author of this website was promised results of the survey made by Graziano Ferrari, but it was repeatedly refused. It was assumed the group intended to publish their own paper and out of courtesy to the publication rights normally afforded academic research, what little information was available to this website was held back.

A small paper was later published in 2015: “Il sistema di vapore delle Piccole Terme di Baia” by Graziano Ferrari, Ivana Guidone & Raffaella Lamagna.

The English abstract includes the following:

“We provided a speleological, archaeological and naturalistic documentation. It resulted in a new survey and an analysis of its morphological and scientific characteristics. The nearest part of the cave was intended as a steam system to serve the surface baths; the farthest areas are more problematic.”

The new survey mentioned has still not been published as I write this in December 2019. It seems one co-author of the above paper has never seen the survey, published or otherwise. One can only presume some kind of conflict among the team that Maura Medri has mentioned.

As for the farthest areas being problematic, the nearest areas must also be problematic, as the steam has to come from the farthest areas.

In substance, Doc Paget’s distance measurements proved to be very accurate, as are the features inside the hill. His sharp bend in South 120 and his extra arches at the Sanctuary are wrong. His assumption about the Dividing of the Ways is slightly wrong.

Taking compass bearings in South 120
Taking compass bearings in South 120

Phase 1 baths

It is suggested that in Phase 1 of Le Piccole Terme the Greek Temple was already there and that the first spa building was the round Tholos and Big D next to it, with some kind of connection too with the Greek Temple.

On the ground there is a significant change in level between the baths and the Tholos, this is not apparent on a simple plan.

An examination of this idea when we add the reality of the tunnels behind reveals several issues for consideration.

  • Firstly the tunnel system is hugely out of all proportion to the very modest bath at the surface.
  • Secondly, to suggest they started to dig into solid rock behind the Tholos and then decided to take a 30º degree detour to the water source is absurd. If you know the water is there, then you take the easiest and shortest route.
  • Thirdly, the evidence surely shows that Phase Zero was a direct passage towards the Greek Temple, which probably stood clear of the cliff before later rockfalls happened. The original entry into the bedrock can be seen inside the tunnel. It was necessary to tunnel and support the way back in, in a later Roman phase.
  • Fourthly, the presence of lamp niches in the tunnels, some 550 of them in total, are not easily explained away as being necessary in a steam channel. Who lit all the lamps and when?
  • Fifthly, the multiple doorways that are to be seen at the steam source. If it is hot enough for people then it is not hot enough for baths. You cannot have it both ways. Doorways mean people.
  • Sixthly, the problem of heat preservation along 170 metres of tunnel. This will be addressed lower down the page.

Phase 1 with the tunnels added to scale
Phase 1 – with the tunnels added

A section Through the Tholos

This slice through the Tholos faithfully records both the pipe at the front under the hypocaust floor and the pipe through the wall into the tunnel behind, inside the cliff. The diagram is based on one from “Baiae-Misenum” by d’Ambrosio, and Borriello, published in 1979.

This diagram is slightly disingenuous, because the inference might be that the heat source is conveniently adjacent to where it is needed.

When we add the tunnel complex, the situation becomes clearer – the distances the hot steam must travel are great, 170 metres. Also there are all those other tunnel walls to heat too.

It is the other way round. An alternative possibility, probably the correct one, is that condensation drained conveniently into the tunnel through the small pipe. The heat inlet is under the front door. The tunnel’s prior use may have been abandoned by this time. A similar pipe with a cemented drainage lip over it exists in the adjacent bath.

Heat Loss

270 has a height of 1.8 metres, is .55 metres wide with a rounded roof. The length to the Dividing of the Ways is 124.5 metres.

290 has a height of 1.8 metres, is .66 metres wide with a rounded top and the distance to the Styx step is 45.7 metres.

These give a combined surface area of 768 square metres of wall to heat up and we’ve not yet entered the bath.

Thermal conductivity, ie heat loss, through solid igneous rock, has a U-value of about 3 – quite high compared to many building materials. U-values are calculated at 75º F, 50% humidity and no wind.

A heating engineer might be able to calculate the power necessary to heat all the walls and the baths, from such a distance.

The areas are conservative – we have not yet taken into consideration any additional factors such as the upper tunnels. 

North and South 120s will add a further 223 square metres of surface area to the above calculation alone. As the surface here is rough hewn and unlined, you can probably at least double that conductive surface area. Even if the walls were smooth, we are now approaching 1,000 square metres of surface area to heat before we get to Le Piccole Terme.

To have any influence at the surface, over a distance of about 170 metres through a tunnel this size, the heat source must have been boiling hot.

Which brings up the question of why there are lamp niches. There are steps for humans down 290. There are door frames at the Styx, Rise, Sanctuary, Back Passage and South 120.

If The Great Antrum was designed to contain and convey heat it then it would be far better to keep the cross section of the tunnel small, to prevent heat loss – a dimension such as we see in the Serino Aqueduct perhaps? Just big enough for a man to carve it through in the first place and perhaps to enter later for maintenance.

If the heat source was very hot, then how could men dig the tunnel and line it with cocciopesto without being boiled and dying, we might wonder.

A Heat Exchanger?


Swipe over the picture, to reveal the Roman tiles.

The construction above the arches within the hill have interlocking Roman tiles to support that which sits above. This is seen both here and in tunnels elsewhere in the hill where the superstructure needs support.

There is no reason why the extension that we see in the open air today is any different from what is inside the cliff space.

A highly-inventive idea is proposed that above the arches in the Greek Temple area there was once some kind of hot water tank. Presumably this would have to be lined with lead, to ensure it was water-tight.

I am sure that if this tank ever existed, the valuable metal would have been stolen, like the exposed butterfly staples holding the Greek blocks together in this area. No evidence for any such tank structure remains today. 

Would there ever be enough heat at the surface left to heat this tank of water before it served the baths? It seems unlikely, 


Could a tank of water have once sat above the trench in the area shown here? Possibly, but a hot tank seems doubtful.