(See first the general description of the Episcopal Ensemble of Terrassa)

San Miguel, located in the middle of the episcopal ensemble, between Santa María and San Pedro, is the best preserved of all three. In principle, it seems unquestionable that it is a baptistry, to such an extent that its first restorer, J. Puig i Cadafalch, installed, not without critics, the baptismal pool, as the original one had disappeared.

From the outside, with an interesting effect of volumes, it appears as a set of roofs, two, three and four pitched, at different levels that makes us think that we are in front of a much more complex building than the structures of three nave basilical plans and an apse that initially constituted Santa María and San Pedro. But it is when we enter the church when we really find ourselves in a totally different world. In fact, its set of arches, upon eight columns with capitals that form the square that supports the central vault, surrounded by four compartments that correspond to the sides of a Greek cross; another four in the corners of the square and the apse, that projects in the eastern side with access through a great horse shoe arch, as well as its complicated system of vaulting of the different spaces, all of that in less than 100m2, offers one of the most impactful images of Spanish pre romanesque art.

The shape of its plan that, as we have mentioned, is a Greek cross inlaid in a square and, with an unattached apse oriented to the east, as in the other churches of Terrassa, clearly links with the set of cruciform Visigothic churches that originated the importation of the model of the mausoleum of Gala Placidia in Rávena for the building of San Fructuoso de Montelios, that, from our point of view, was followed in churches like Santa Comba de Bande, San Pedro de la Mata or Santa María de Melque and that would continue in Carolingian constructions like Geminy-des-Prés and even in Santa María de Lebeña, already in Mozarabic times, and in the mosque of El Cristo de La Luz in Toledo, completed in 999.

However, San Miguel is the only one among them that maintains the arches’ structure upon columns, supporting the central vault that we see in Montelios, altough the addition of the compartments in the angles improved the stability of the ensemble. In this case, instead of the three arches on each side of the square that we see in San Fructuoso, there are only two by side, leaning upon for columns in the corners and another four, thinner, one in the middle of each side of the square. Its covering system, that possibly had gone through modifications regarding the original structure, presents very interesting solutions, as it is based in four groin vaults, in brick, in the four arms of the cross, another four semi domed vaults in stone, in the four compartments that form the angles of the square, so that, leaning the curved part upon the external walls, they reinforce the supports of the central vault, hemispherical, upon squinches, that have four windows, one at each side of the dome. To complete the stability of the ensemble, in the separation between the arms of the cross and the lateral compartments, there are eight rib arches that start form the four angles’ columns and lean in the external walls, reinforced in the contact points by pillasters in the inside and buttresses in the outside of the church. It is necessary to bear in mind that the arches that form the central square and support the vault are elevated round arches; different to San Fructuoso, where the arches are shoe arches. However, San Miguel, reminds us internally, maybe for the sensation of lightness, and externally for the effect of volumes of its roofs, much more the Mausoleum of Montelios than the Carolingian art, which in general is much heavier.

A special mention deserves its apse to which access is gained by a continuous horse shoe shaped arch. It is placed in the eastern side, forming an external compartment to the square with inner horse shoe shape, very similar to the one in santa María. However, externally it is octogonal, something without any precedents in the architecture of that period in Spain, except for the remains of a Visigothic church discovered in the cloister of the monastery of San Cugat del Vallés, of the same shape internally and externally, although in this case it is five sided, so it could be a local variation of that period.

The building had three doors, one at each side except in the apse. Today only the one at the southern side is left, much modified and possibly it let the direct access to the northern nave of Santa María.

Another interesting matter is the existence of a martyrial crypt located beneath the apse, to which access is gained by staircases at the southern side of its interior. It has an aisle, parallel to the eastern wall, and a perpendicular chevet, with a plan formed by three horse shoe shaped lobes, very similar to San Pedro’s, although in this case the covering of the lobes is achieved by a barrel vault and the central square’s is flat.

In the western facade, in an area with a different building style, of worse quality, we find a big very pronounced horse shoe window, clearly Mozarabic, a proof of a first partial reconstruction around the 10th century, previous to the Romanesque one, but also showing that the original ensemble is prior to the Mozarabic period.

With regard to the decoration, besides the existent painting in the apse, there are only left in the capitals of the inner square, schematized Corynthian and little homogeneous, probably reutilized from the Roman period, as well as the columns that are of different thickness. They support the set of arches that on their turn support the walls of the central vault.

As can be seen in everything mentioned above, San Miguel de Terrassa is one of the most exciting and conflicting monuments of all Spanish pre romanesque art. For its building quality and design, both internally and externally, is comparable to the best European monuments built after the decay of the Roman Empire and previous to 900, with the added value that it forms part of an homogeneous ensemble with regard to an integral planification and an initial construction type, utilizing the same bond, similar and undoubtly of the same period. On the other hand, both, its structure as well as its architectural solutions have precedents in Spain prior to the 8th century, but they also extend in time to some Carolingian constructions that are considered to have been influenced by the Visigothic art. As a compounded complication we find that the three buildings suffered continuous modifications and additions since they were built until the 12th century, and no archeological studies have been accomplished that cover the whole ensemble; besides, there is very little published information about those that have been attained partially.

For all that has been mentioned above, it is practically impossible to conclude a definite dating, although there are likely options: either its initial construction took place during the zenith of the bishopry of Egara in the last Visigothic period, at the end of the 7th or beginning of the 8th centuries, or that they had been modified during the settlements of the Hispanic March in the second

half of the 9th century. Therefore, although without total certainty, our impression is that its origin is Visigothic. We shall limit ourselves to put forward the points we believe most meaningful in favour of each of these two options: