‘In 1821 I was required by His Majesty’s Commissioners to make designs for an additional church, which was proposed to be erected in the parish of St Marylebone, the expense of which was to be defrayed partly by the parish, and partly out of the parliamentary grant.

I made designs accordingly, suitable, as I considered, in external character and internal decoration, to the accommodation of the inhabitants of that extensive and opulent parish; but as the parish subsequently declined to make the necessary advance for the purpose, these designs were relinquished, and others of a less expensive character made, which were afterwards carried into effect under my superintendence, and paid for wholly out of the parliamentary grant, excepting a small sum subscribed by individual parishioners for the enlargement of the steeple, which had been reduced in order to keep the expense within the limits prescribed by His Majesty’s Commissioners wrote Soane in his Memoirs of the Professional Life of an Architect, 1835.

There are two references to the ‘Model of a Church’ in Sir John Soane’s Notebooks. The first is in the entry for February 6th 1822: ‘...Office of Works to meet Laing and P--- / Mr John Soane called, qy [=query] this day or the next (some time after I went out) / Model of Church’. The second relevant entry is for April 3rd 1822: ‘Miss Levick at the Bank 12 o'cl; examine Model of Church…’.

There is a further reference in Soane Archives (Box 8 set XLVI / 41) which reads: ‘Account of Time & Materials to Model of Church charged in the Carpenters Accounts by mistake Ladyday Qr 1822 & deducted from that Acct (see Abstract Lady Day & Midsr 1822) and Harrisons Book 31 Janr 1822’. The model therefore had been made by March 1822 and had apparently been charged to the wrong account.

As we know from drawings for Holy Trinity that Soane was working on designs similar to model MR79 in early 1822, it seems likely that these references relate to it. Yet Bolton’s Works of Sir John Soane includes the transcription of the account: ‘Feby. 1824 – No 3 Making a Design for a Church proposed to be built on the North side of the New Road, making various fair drawings of same, … , a Model, and an Estimate of this Design as approved by His Majesty’s Commissrs.’ The model cost £25:0:0 and was also for Holy Trinity. Soane’s designs of 1824, however, have little in common with the design as shown in MR79, and we must assume that a second Model, now lost, was made in that year.

The Model MR79’s roof can be removed in sections, so revealing Soane’s design for the interior. Alternative treatments are given to the galleries on the N and S sides. To the S the nave and aisles are separated by a row of five round-headed arches on slender piers, with stepped gallery behind. This design first appears in drawing 54/2/8, View of the Interior looking towards the Altar dated May 1820,then in drawing 54/1/27 dated 1822. On the N side, the gallery has a curved front that breaks forward between the piers, above a row of segmental arches with dropped keystones. Segmental arches feature in various drawings (minus dropped keystones), including 54/1/20 of November 1820. In the interior as shown in the model, the aisle arches are mirrored by blind arcading on the outside walls, an element not found in any drawing. In the final, executed design Doric columns support the gallery and the piers to the arches at gallery level.

Holy Trinity’s main entrance on the model is through a vestibule at the W end, fronted by four sturdy baseless Doric columns (they are not in antis as is the Ionic order of the executed design). Two further pairs of Doric columns, of identical proportion, mark the way through the vestibule to the church. The vestibule has apsidal ends to N and S with doors leading to the stairs in the corners of the building. The first drawing to illustrate this design is dated November 1820 (54/1/8) and perspectives of something very like appear on sheet 54/1/22, also dated 1820. Although Soane was to work on many alternative designs for the W entrance, he returned – if briefly – to this solution in 1822, as can be seen in sheet 54/1/7.

The other entrance to Holy Trinity was at the E end. In the model, steps lead to a vestibule behind a pair of baseless Doric columns, with doors to N and S. They lead to narrow halls with apsidal ends that, judging from drawings, were to have held stairs (in the design published in Soane’s Designs for Public and Private Buildings of 1838, these rooms are differently orientated and run along the E wall). The halls give onto a very square vestry and robing room to N and S. This design does not appear frequently in Soane’s drawings. In sheet 54/1/3, dated 1822, the entrance has been given two flanking columns and Soane is working on the location of the narrow halls; the SE corner is closest to the model design. In 54/1/4, also dated 1822, the proportions of the vestry etc. are close to those of the model if you follow the amended design traced out in red ink. Elevation to the East Front (54/1/30) of 1822 features two alternative designs: the design to the right is the same as that on the model, including the Diocletian window in E wall above the altar.

Even the earliest drawings for the side elevations are close to the model design: 9 bays, paired Doric columns with a window between in the second bay to E and W ends, the use of inset round-headed windows, a clerestory of 5 bays with segmental windows. These are all characteristic of the Flank Elevation Design No 1 of November 1820 for instance. But the drawing that bears the closest resemblance is sheet 54/3/2, dated 15th March 1822. It differs from the model in that pilasters have been placed between the clerestory windows, a Diocletian window appears at the W end of the clerestory, and the steeple is of other design.

The steeple in fact appears is innumerable forms in the drawings for Holy Trinity. With the exception of its rusticated base, the lower stage of the model’s steeple is the same as that executed. The only drawing with any likeness to the second stage is sheet 54/1/32 (1822) but there are major variations: the faces of the drawn polygon are of equal size and have segmental pediments, not alternating segmental and triangular pediments.

This model is clearly a re-working of elements designed as early as 1820. It is certainly quite unlike the executed design, approved by the church Commissioners on July 13th 1824 (see Soane’s Notebooks). The church’s working drawings were not made until over a year later – the earliest is dated August 1825 – a delay that had been caused by problems over the site.