The ‘Poole taile of Hayemill’ is referred to in the 1495 report on the boundaries of Yardley, and Hay Mill ‘Pool Tail’ is referred to in the 1609 report. It is likely to have been early medieval in origin, the property of Hay Hall: the delaHaye family dwelt there from the late thirteenth century, and the estate counted as a sub-manor in Yardley from the fifteenth century. The mill is shown on Beighton's map of 1725, powered by a long leat from the Cole and by the Tyseley Brook (River Lee). A postmill appears on Red Hill nearby, and was perhaps a replacement for Hay Mill when it converted to grinding.
In 1795 the pool dam of Hay Mill was cut down by persons unknown (Aris’s Birmingham Gazette). In 1820 the mill was described as a blade-mill, suitable for corn-grinding or paper-making. The millhouse was apparently a separate building, built against the dam as was the mill itself.
There were two rebuildings of the mill after 1800. The first was completed by 1817, and was on a site two hundred yards north of the blade-mill. The earlier (seventeenth century?) mill survived into the twentieth century. If both earlier and later mills were at work together, then ‘Hay Mills’ is correct, and not due solely to the local habit of pluralizing names. The old pool was retained, and a larger one of about three acres was made below. Between 1836 and 1840, William Deakin, swordsmith and gunmaker, operated the mill, or had it working for him, on his arms contracts.
The partnership of Smith & Horsfall was in business at Hay Mill from 1841 to 1846, and the following year James Horsfall began his long career. Soon afterwards the mill was rebuilt or very considerably extended along the dam.
In 1860 wire-drawing machinery from Penns Hall Mill north of the Tame was installed at Hay Mill. Webster & Horsfall obtained contracts for iron sheathing wire for the first Mediterranean and the second (successful) Atlantic Cable, and further drastic reconstruction took place in 1865. Continuing faith in water-power had been shown in the earlier rebuildings, but it was now abandoned, the pools were drained, and the factory was extended along the side-race, the tail-race being infilled. In 1863 Horsfall erected a school-room at the lower end of the factory, and ten years later a church, St. Cyprian’s, which some fancy moved him to build over the race: the culvert curves from the south end to the west side of the church, and two millstones may be seen in the water there.
In the 1920s Tyseley Salvage Works and Destructor were built over the site of the early mill, cottage, and pool, between canal and race: the latter survives because it assists the flow of flood-water from the confluent Cole and Spark Brook beneath the canal and railway embankments. The former low meadow between river and race has been transformed by the deposit of clinker from the Destructor into a plateau as high as the embankments, so that the watercourses are now deep gorges, bush-clad and dramatic.
When Tyseley Destructor Works was replaced by the great Waste Disposal Unit in the 1980s, the surviving upper end of the lower and later pool was enlarged to make an attractive quadrilateral pond. It is still fed by the original head-race, which goes under the railway, the canal, and the concrete platform of the W.D.U., and is drained by the long tail-race beside the wire works. Both sides of the valley have been drastically altered by works associated with the Small Heath Bypass, the ‘Ackers’ leisure park, the W.D.U. and the provision of a balancing lake above it.
(More recent notes on the watercourses appeared in 'All around the 'Ackers'; these are reproduced here (ed.)):
Field Drawing Sheets (1:126720) for the Ordnance Survey, 1817, show the Cole, the Spark and the Lee joining in a pool south of the canal embankment, and a single channel flowing from there. At Hay Mill Cottage, a headrace continues to a triangular millpool, and the main flow curves round it; the two arms rejoin 400m south of the Coventry Road.
The First Edition of the O.S. One-Inch Map (1:63360) c.1822 is more accurate in some respects and may be more reliable hereabout. It shows three channels under the canal - Spark, Cole and Lee; the last feeds a rectangular pond above the embankment and serves as the headrace to Hay Mill. No confluence of channels is shown above the mill, and the tailrace continues to beyond the Coventry Road. Doubtless in flood the three courses combined. The reason for the long tailrace is that after turning the breast-shot wheels, the water was in a well below river level, and could be returned to it only by a channel of lesser gradient: by trial and error this had been found to require a leat of 400m, with a culvert under the turnpike.
The 1841 Tithe Map of Yardley (1:5280) shows two channels only. The Spark is almost joined by the Cole and a straight branch from the latter is fed by the Lee, having three pools upon it and a floodrace at Mill Cottage running beside the lowest pool to join the Spark below the mill. This finely-drawn map is the first to achieve modern standards of accuracy. On it we see that the Cole flow is normally down the headrace: the channel leading towards the Spark is a 'floating course', a dead-end cut which would overflow into the brook only in flood. Downstream of the canal, a channel leads to the floodrace from the embankment foot, and this may indicate a culvert used to drain the meadow on the south side of the bank.
Blood's splendid map of 1857 shows the Oxford Railway viaduct over the canal and the embankment across the confluence meadows south of it, cutting in two the uppermost of the three pools. We see also that for reasons unknown and not readily discoverable, the combined Spark/Cole channel (the floating course having been extended) has been diverted from the old Spark course to flow for 120m between the two banks before going under the aqueduct from which the Park takes its colloquial name. There is no doubt about this diversion: on all subsequent maps the ancient boundary is drawn following the old brook course directly across the canal. No trace of the original 1790s culvert on this line can now be found.
Floods were a greater problem when two barriers impeded flow down the valley, and by 1887 (O.S. Map Second Edition One-Inch) and First Edition Six-Inch (1:10560) a relieving course had been made. A weir just below the junction of the Cole and the Lee fed water to this: the brick-walled channel between the two banks joins the Cole/Spark at the aqueduct: this was given two arches, one of which is normally dry.
Widening of the rail embankment to support extra lines after 1907 necessitated extension of the three culverts beneath it: as by then water was no longer used for power at Hay Mill, a low arch was made for the headrace, sufficient only to prevent the channels from drying up.
Severn-Trent Water Authority has acquired control of the Cole system from the City and Trent Water Board. Amenity is desirable, but flood control is always the chief concern. A double concrete weir has been built below the confluence of the Cole and the Lee: the higher upper dam maintains the river level for the benefit of wildlife and citizens, and the lower forms an intermediate basin which prevents erosion of the soft clay bed by water which would otherwise fall some 1.5m. The overflow channel, with a newish concrete portal in the rail bank, has no approach arm at the weir: in exceptional floods, water can drain into a ditch parallel to the bank and so into the channel, but usually the only flow therein is what has backed up from the main Cole/Spark course. Downstream of the Waste Disposal Unit, a balancing lake has been created: excess water can be diverted into this and released when the river is low.
More recent maps were also produced for 'All around the Ackers', so the original map has not been included here.
Provisional list of Cole valley watermills
Peterbrook, Dobbs, Crab, Kilcop and Forshaw Mills
Colebrook Priory and Old Mills
Broomhall and Lady Mills
Possible mills in Greet and Tyseley, Medley's Mill