Exploring the Relics of Country Mills in Connecticut

 Jamie H. Eves

A stiff June breeze blew through the northeast Connecticut hills, skimming the surface of the beaver pond and scattering some of the sultry summer haze that had settled over the oak and beech forest. Silvery tree trunks stood like gaunt skeletons in the bright blue water, evidence that the pond was fairly new. A steep grassy bank rose near the outlet, littered with the pointed, gnawed stumps of beech trees. Groves of living beech grew further up the bank, the light green leaves purring softly in the wind. Nearby, Columbia Lake Brook drained noisily from one end of the pond, tumbling over two beaver dams and dropping into a shady ravine that carried it out of the hills and into the broad valley below. I clamped my hat more firmly on my head, swatted a mosquito, and tromped eagerly down the ravine. I was hunting for old mill dams.

I had heard that perhaps as many as five country mill dams had once spanned a three-quarter-mile stretch of Columbia Lake Brook in Columbia, Connecticut. Today the land around the brook is woodland, second- and third-growth forest owned by the Joshua’s Tract Conservation and Historic Trust, founded in 1966 to preserve the rural character of various parcels of land in northeastern Connecticut. The Trust’s holdings are like a giant outdoor museum, filled with evidence of nineteenth-century country life. Miles of relic stonewalls indicated that in the 1800s most of this site had been farmland, a patchwork of hay fields, maize fields, pastures, wood lots, and dirt roads, now abandoned and grown up into forest. (To learn more about preindustrial country mills in New England in general, click here.)

Columbia Lake Brook flows north from the beaver pond (which apparently existed in the early 1800s, reverted to swamp when the beavers left, and became a pond again when they returned in the late 1980s) for about three-quarters of a mile before entering Columbia Lake, a reservoir constructed in 1865 to provide water for the turbines of the American Thread Company’s big textile factory in Willimantic, six miles away. According to the Trust’s Walk Book, the relics of four or five country dams lay along the brook. I wanted to measure and map them.

U. S. Geological Survey map, 1890, Gilead quadrangle, showing Columbia, CT, upper right. Columbia Lake is quite visible on the map. It is an artificial lake (as are most lakes in Connecticut), created in the late 1800s by the Willimantic Linen Company as an upstream reservoir for its water-powered textile factories in Willimantic, CT, a few miles downstream. Columbia Lake Brook flows into the lake from the south. The map, like as USGS maps, is public domain. This copy is in the collection of the Perry-Casteneda Library at the University of Texas.

Unfortunately, the two or three upper dams now lay under the rising waters of the beaver pond. These had probably been only water-storage dams. Downstream in the ravine, however, I found the relics of the other two dams, as well as the foundation of one of the mill houses. The upper of the two dams was about 10-15 feet high, nearly 15 feet thick, and constructed of stone covered on top with earth. A layer of earth also covered the upstream side of the dam, probably sediment from the former millpond, which had long since drained through a breach in the dam. An old country road on the west bank, long since abandoned, had once provided access.

The mill house was about 180 feet downstream – a drop of 25 or 30 feet from the top of the dam, enough to power an efficient overshot wheel. Although the brook was only four or five feet wide and two feet deep, it would have provided ample power for a country gristmill or sawmill. The foundation of the mill house was built snugly into the stream bank and made of stone. The walls were two to two-and-a-half feet thick. The foundation measured 30 feet by 24.5 feet. About two-thirds of it had been covered by the now-vanished wooden superstructure of the mill house, with the other third jutting into the brook as a pier to hold the waterwheel. The wheel well was seven feet wide, large enough to hold an overshot wheel. The under story of the mill was only four feet high and had a stone floor, meaning that the bulky mill machinery must have been located on an upper level. Stone piers between the dam and the mill house indicated the likely location of a wooden raceway, which would have carried water from the millpond to the waterwheel. Wooden flash boards atop the dam (there was no evidence of a sluice gate) would have permitted the miller to send water down the flume to the waterwheel. If the water level in the millpond dropped too low, the miller could have ascended to one of the water-control dams upstream to release more water.

Following the old road, I headed further downstream. After about 300 feet I found the relics of the second dam. This one, too, had been constructed of stone and earth. It was about 15 feet high and 13 feet thick. I found no evidence of a mill house, but that didn’t mean that there hadn’t once been one – the relics may have been removed by farmers looking for boards and stones. Several hundred yards further downstream the road crossed the brook via an old stone bridge. Continuing between two crumbling stonewalls, it soon opened onto farm fields not far from Columbia Lake.

Two days later I stopped by Columbia’s Saxton B. Little Library, where I met town historian De Ramm. With her help, I unearthed two nineteenth-century maps of Columbia. A copy of an 1857 wall map, made before Columbia Lake was constructed, clearly showed the brook (then apparently nameless), along with one mill dam, a millpond, a sawmill just below the dam, and – further upstream – a beaver pond. The map also showed the ravine, but no roadway, which must not have been a public road. The mill was labeled “Holbrook’s sawmill.” A second map, from a county atlas published after Columbia Lake was constructed in 1865, also showed one mill – an “s. mill” – but did not give the miller’s name. It seemed unlikely that that there were ever as many as five mills on the stream.

My next trip was to the State Library in Hartford, to look at federal census records. Beginning in 1850 the census recorded information on all of the “industries” – including country mills – in each town in the United States. The town-by-town reports are available on microfilm. With my head thrust into a microfilm reader, squinting to make out the enumerators’ handwriting, I scanned the 1850, 1860, 1870, and 1880 reports for Columbia, hoping to find out more about the mills on Columbia Lake Brook. The 1850 census showed only one sawmill and one shingle mill in all of Columbia, both belonging to Joseph Clarke. Other evidence indicated that Clarke’s mills were in a different part of town and not on Columbia Lake Brook. The 1860 census, however, showed eight sawmills and two shingle mills, a remarkable increase in just 10 years. Clarke’s two mills were still listed – but so, too, was a sawmill owned by L. Holbrook. I had found it!

Holbrook’s mill was typical.  It had a capital value of only $500, about average for a country mill. It had only one saw and one employee, probably Holbrook himself. It was a custom mill, taking in 80 cords of logs each year from neighborhood farmers, and producing 80,000 board feet of lumber valued at $800.

The 1870 census showed seven sawmills and/or shingle mills (it did not distinguish between them) in Columbia, but none belonging to Holbrook. Had the mill closed? Had it been sold to someone else? It was impossible to tell. The 1880 census showed only one water-powered sawmill or shingle mill in all Columbia – Joseph Clarke’s. A shift in technology had occurred. Four efficient new steam-powered sawmills were now in operation, the census said, driving the old water-powered mills out of business. Clarke’s mill was clearly out-classed. In 1850 Clarke’s sawmill had produced $550 worth of lumber and his shingle mill another $525 worth of shingles; in 1860 his combined output was valued at $1,000; and in 1870 he produced $2,500 worth of lumber and shingles. But in 1880 his combined produce was valued at only $450. The Age of Water was coming to an abrupt end. The Age of Steam was replacing it.

Historians sometimes refer to the period of American history before the Civil War as the Age of Water or the Age of Wood. And so it was. Water provided power. Wood provided fuel, building material, fence posts, ships, wagons, tools, and other vital goods. When used for building, wood comprised the flesh, muscle, and sinew of the structure. But underneath the wooden exterior there usually lay a skeleton of stone. Time has destroyed much of the old wood superstructures, but the stone frequently remains, like the mill dams of Columbia Lake Brook, buried beneath the earth, leaves, and brambles, reminders of an earlier time.