Evidence of a prehistoric two-weir fish trap at at China Lake Stream, Sebasticook River, Winslow, Maine.

By Douglas Watts
Augusta, Maine
September, 2009.

Evidence since the removal of the Fort Halifax Dam in 2008 suggests the presence of a prehistoric two weir fish trap for alewives and/or American eel at China Lake Stream, Sebasticook River, Winslow, Maine.

General Locality Description:

China Lake Stream (also known as "Mile Brook" or "Outlet Stream") flows north from its source, China Lake, located in China and Vassalboro, Maine to the Sebasticook River in Winslow, Maine, one mile above the confluence of the Sebasticook and Kennebec Rivers. The stream is approx. 6 miles long.

Until 2008, the lowermost five miles of the Sebasticook River and the lowermost one mile of China Lake Stream were flooded and impounded by the Fort Halifax Dam, constructed by Central Maine Power as a hydroelectric dam in 1908. During this century-long period, the lowermost mile of China Lake Stream was a shallow and narrow pond-like impoundment with a width of 100-300 yards and a substrate of deep organic muck.

View Larger Map
In this satellite image, the Kennebec River flows from north to south on the far left, is joined by the Sebasticook River, flowing from the east to west, with China Lake Stream entering the Sebasticook River from south to north in the right center. This image was taken before the Fort Halifax Dam at the mouth of the Sebasticook was removed in 2008 and shows the dam's impoundment extending up the Sebasticook River and China Lake Stream. The Fort Halifax dam can be seen as the large, bright feature in the Sebasticook just before it enters the Kennebec River.

---

In summer 2008, the Fort Halifax Dam was removed by its owner, FPL Energy. Since the dam's removal, the flow of China Lake Stream has downcut through a century's accumulation of soft silt and sediment and the original channel has largely been restored to its natural gravel, cobble and boulder substrate. Thick (4-8 feet) beds of accumulated silt and sediment still cover the former impoundment on both sides of the newly incised channel. During the past two growing seasons these areas of rich silt and muck have dried out and have become thickly vegetated with grasses and other pioneer annual plants. This thick silt bed completely blankets and obscures the original soil surface along the first one mile of the stream.

This photo shows all of the changes caused by the impoundment of China Lake Stream by the Fort Halifax Dam since 1908 and its removal in 2008. The tree line in the background was the top of the impoundment, the grass line was the bottom of the impoundment, the stream channel shows the original channel of the stream downcutting through the silt and muck of the former impoundment, and the ledge of tightly folded Waterville slate in the foreground shows the natural hydraulic control of the streambed.

This photo shows all of the real important changes on China Lake Stream. Dan Watts and Queegueg T. Dog, Ph.D. wade through the rushing waters of China Lake Stream on July 4, 2008, where no child has been able to wade since the 1800s.

This photo shows the five foot thick bed of sawdust, muck, silt, trash and woodchips that has accumulated in the bed of the formerly impounded reach of China Lake Stream during the past 150 years. The natural shoreline of the stream is buried somewhere below the base of this stack of recent debris.

---

A point approx. 1/4 mile below where State Route 137 crosses China Lake Stream marks the upstream limit of the former Fort Halifax Dam impoundment. This point is approx. one mile above the stream's mouth and is marked by a mature forest (primarily hemlock) growing directly alongside the stream with no areas of impoundment inundation and sedimentation. The stone weir sites are several hundred yards below this point at the uppermost extreme of the impounding effects of the former Fort Halifax Dam. Based upon the vertical height of the "bathtub ring" along the wooded margin of the stream, it appears the site of the stone weirs was impounded by the Fort Halifax Dam by about two feet of water.

Site Description:


View Larger Map
This satellite image shows the weir sites at the bottom left, with China Lake Stream flowing from south to north, (bottom to top). This image was taken before the Fort Halifax Dam was removed in 2008 and shows the extent of the impoundment in China Lake Stream as well as the course of the stream's original channel, which has now been restored.

---

The weir site consists of the remains of two barriers made of river stone that perpendicularly traverse the channel of China Lake Stream. The two stone barriers are approximately 250 feet apart in a stream section that is especially shallow, flat bottomed and composed of golf-ball to baseball-sized cobble. The upper barrier is a 1-2 foot high "wall" of bread loaf sized river stones that completely crosses the stream except for a 3 foot wide "chute" hard on the left hand side of the stream (looking upstream). The barrier curves slightly downstream. The barrier is quite obviously manmade.

Upper weir site, low water, Sept. 8, 2009.

Upper weir site, high water, July 4, 2009.

---

The lower weir feature is much more fragmentary and incomplete, but in some ways is more diagnostic. It is located just upstream of a small riffle feature in the stream and at the tail of a 250 foot long, very shallow pool. At low stream flow (as observed in early Sept. 2009), approx. two thirds of the stream channel (starting from the right hand side) is dry and composed of small, evenly sized cobble. This flat cobble bar focusses the entire flow of the stream hard on the left hand bank of the stream channel, where the bank quickly climbs about five feet into a grove of mature hemlock.

Queequeg T. Dog, Ph.D. at lower weir site.

---

The presence of a stone weir at this lower site is conjectured because of the anomalous presence of a 3-foot wide line of large stones crossing the stream perpendicularly in the middle of the cobble gravel bar. These large stones are set deeply into the smaller cobble bar and form a roughly straight line for a distance of approx. 10 feet across the middle of the stream channel. This line of stones can only be observed at very low water. The orientation and location of the stones cannot be explained by natural sorting and scattering by the stream's flow at this location. A number of large (basketball sized) rocks are anchored hard on the left bank in a line fairly consistent with the mid-channel row of stones.

Other Evidence of Prehistoric Use and Habitation

Supporting evidence for these stone features being weir remnants is provided by two accumulations of fire cracked rock (FCR) and shards of worked Kineo rhyolite on opposite banks of the stream directly adjacent to the site of the lower weir. Extensive searches of both sides of the stream bank for 1/4 mile above and below the lower weir site during three site visits in the past year have failed to locate any FCR or flint/chert debitage. FCR quickly disappears 100 feet above and below the lower weir site on both sides of the stream bank. All of the debitage found to date is from Kineo rhyolite and is only in large rough shards, most with one side being the natural rounded surface of a glacially rounded cobble. FCR fragments outnumber debitage by about 30:1 and are most densely concentrated on the left hand side of the stream bank approx. 30 feet below the lower weir site. While shards of 19th century, glazed redware pottery are abundant in the streambed and shoreline above the site to the Route 137 crossing, no shards of prehistoric ceramic have been found in the area of the conjectured weirs.

Kineo rhyolite shard from the left hand bank of lower weir site.

Kineo rhyolite shard from the right hand bank of lower weir site.

---

Examination of the soil surface at the stream bank at the lower weir site shows a clear horizontal interface of hard marine clay overlain by a darker, browner, less cohesive silt layer, finally overtopped in places by a layer of finely ground sawdust, sawn wood chip fragments and woody debris, presumably deposited by 19th century sawmills (there is a breached 19th century dam approx. 1/4 mile above the weir site).

Queequeg T. Dog, Ph.D. and a stick show the horizontal interface of the 20th century Fort Halifax Dam impoundment layer of silt and original marine clay shoreline of China Lake Stream just below the lower stone fish weir. Fire cracked rock (FCR) and worked shards of Kineo rhyolite are found in the area below the "silt line," which is the natural surface of stream bank.

---

Are the conjectured weir sites the remains of 19th century dams or 20th century "swimming holes"?

Due to their placement, orientation and design, the two in-stream features are clearly manmade. This does not necessarily mean they were made by prehistoric people as fishing weirs. These features could have been made by European settlers. A number of independent lines of evidence rule out this possibility.

The construction of the Fort Halifax Dam in 1908 impounded and flooded the site of the two weirs by at least several feet. This eliminates the chance that the barriers were built in the 20th century by local kids as a "swimming hole." But what about a 19th century swimming hole? This can be discounted for two reasons. First, the stream reaches above either barrier are far too small and shallow to be worthy of such effort to make a 19th century swimming hole. Second, the extensive use of China Lake Stream as a conduit for industrial waste and household sewage during the 19th century would have made the stream less than desirable as a swimming hole for local youth.

Are these features the remnants of 18th or 19th century mill dams?

Wooden mill dam remnants are exceptionally well preserved if kept underwater. The sites of the stone weir features have been constantly underwater since the construction of the Fort Halifax Dam in 1908. Any timber remnants in the riverbed from a 19th century mill dam would still be well-preserved and visible in the streambed today. No timber remnants are visible in or along the streambed at the two weir sites.

A second clue is the presence of a very old, well-preserved timber dam on China Lake Stream one half mile downstream from the weir site, just above the Garland Road bridge. This dam is obviously of 19th or early 18th century vintage because it was inundated by nearly ten feet of water by the 1908 Fort Halifax dam impoundment. The design of this small dam is identical to timber dams built at the mouths of Riggs Brook, Goffs Brook and Seven Mile Stream along the Kennebec River in Augusta, Sidney and Vassalboro in the late 1700s and early 1800s. The latest date of construction of these small timber dams is the 1830s -- before the Augusta dam was built on the Kennebec River in 1837, which flooded and inundated these dams and preserved their timbers until the Edwards Dam was removed in 1999 and these dams were rediscovered.

A last clue is that by the mid 19th century, mill dams on rivers and pond outlets in the area were constructed with large, roughly rectangular pieces of quarried stone to increase hydraulic head and to reduce the chance of dam failure during freshets (Watts 2004). An example of this type of dam is found directly below the Route 137 crossing of the stream. Dam sites of this construction type are easily identified by their well-fortified abutments of cut stone in the stream bank, a well-defined "mill pond" above the dam site and the presence of large, square or rectangular stones in the streambed and banks directly below the dam site.

The weir sites lack any of the typical remnant features of a mid-to-late 19th century stone mill dam or a late 18th or early 19th century timber dam. This negative finding is strengthened by the obvious and well-preserved remnants of both types of dam immediately upstream and downstream of the weir site.

A hefty slab of Kineo rhyolite found on the bank of China Lake Stream, still showing the freshly broken surface from when it was shattered 4,000 years ago to make a spear point. In the background of the photo you can see the site of the lower stone weir (at the bend to the left) and in the far background the upper stone weir. Prior to dam construction, more than 500,000 alewives swam past this spot on China Lake Stream every spring to spawn in China Lake, and thousands of 3-4 foot long American eels migrated downstream from China Lake past this spot each fall.

---

Historic accounts of stone weirs for alewives

Pory (1622) and Josselyn (1647) describe the use of piles of stones in a stream to block migrating alewives to aid in their capture:

Pory writes: "In April and May come up another kind of fish which they call herring or old wives in infinite schools, into a small river running under town, and so into a great pond or lake of a mile broad, where they cast their spawn, the water of the said river being in many places not above half a foot deep. Yea, when a heap of stones is reared up against them a foot high above the water, they leap and tumble over and will not be beaten back with crudgels."

Josselyn writes: "The Alewife is like a herrin, but has a bigger bellie therefore called an Alewife, they come in the end of April into fresh Rivers and Ponds; there hath been taken in two hours by two men without any Weyre at all, saving a few stones to stop the passage of the River, above ten thousand."

Hanson (1852) describes an 18th century female settler dipnetting large numbers of alewives from Worromontogus Stream in Pittston, Maine, 18 miles south of China Lake Stream:

"It is related that alewives were so plentiful there at the time the country was settled, that bears, and later swine, fed on them in the water. They were crowded ashore by the thousands. Mrs. David Philbrook, who was a McCausland, was very much in want of a spinning wheel. One day she took a dip net, and caught seven barrels of alewives in the Togus, and took two barrels in a canoe, and paddled them down to Mr. Winslow's, and exchanged them for a wheel."

See Watts (2003) for further 18th century documents. (pdf here)

These 17th and 18th century accounts show that, in a small enough alewife stream, nothing more than a low pile of stones in a shallow riffle would create a sufficient barrier to allow for large amounts of alewives to be taken by nets and baskets. In contrast, the China Lake Stream weir sites show a more industrious intent. First, there are two weirs within a few hundred feet of each other (why not just one?). Second, the size of the structures and the time and effort taken to make them shows more than just a casual, low intensity use and purpose.

A large stone fish weir on Seboeis Stream, Piscataquis River, Howland, Maine. Photographed by Douglas Watts in extremely low water in October, 2002. This weir is still periodically used by local commercial eel fishermen to catch adult eels migrating downstream from Seboeis Lake, which accounts for its excellent condition. Given that the inverted "V" of the weir points upstream, it was most likely first built to capture fish migrating upstream (ie. alewives), which once went upstream to Seboeis Lake and other ponds in the watershed in huge numbers until the runs were extirpated by dam construction on the Penobscot and Piscataquis Rivers in the 19th century. Despite this design favoring the capture of upstream migrants, the weir was mostly likely also used to catch eels migrating downstream in the fall.

---

Using two weirs to create a more effective fish trap:
a design hypothesis.

The location of two prehistoric stone weir structures less than 300 feet apart across the bed of China Lake Stream suggests the possibility these two weirs were built and used in tandem to create a highly effective fish trap to catch adult alewives (Alosa pseudoharengus) migrating from the Atlantic Ocean to their spawning grounds in China Lake.

Recent excavations at the Cates Farm archaeological site at the outlet of China Lake have found alewife bones, confirming the harvest and consumption of alewives by prehistoric people at China Lake. These alewives reached China Lake by migrating from the Atlantic Ocean to the Kennebec River, up the Sebasticook River and up China Lake Stream. Modern biological studies on alewife productivity and the size of China Lake suggests the historic size of the China Lake alewife run exceeded 500,000 adult alewives each spring. Based on modern alewife runs on the Sebasticook River, the alewife run on China Lake Stream began in late April and ended in early June.

A large alewife run forced to migrate up a stream as small, shallow and narrow as China Lake Stream to reach their spawning grounds in China Lake would have created an ideal seasonal fish harvesting opportunity for prehistoric people. The weir sites located on lower China Lake Stream have the prerequisites of a preferred alewife harvesting locality. At the weir sites, China Lake Stream is very shallow (1-3 feet deep), yet narrow (less than 50 feet wide), and is only one mile above the stream's confluence with the Sebasticook River, making it easily accessible by canoe or foot from the Sebasticook, which served as a key travel highway between the Kennebec and Penobscot River drainages. It may not be coincidental that the two largest concentrations of extant prehistoric lithic artifacts on the Sebasticook and lower Kennebec Rivers is found on the Sebasticook River directly across from the mouth of China Lake Stream and a quarter mile below it.

Adult alewives have a powerful homing instinct which commits them to swimming tens or hundreds of miles to reach their natal birthplace in a specific spawning pond. To reach these ponds and spawn, alewives will spend days and weeks trying to swim over or around any obstacle in their way, be it a human-made dam, a beaver dam, a natural waterfall or a log jam in a stream.

A stone and brush fish weir is made by erecting a wall of stones across the stream bed and driving saplings, sticks and brush in between the interstices of the stones to create barrier against fish migrating upstream (alewives, shad) or downstream (adult American eel). The advantage of a stone and brush weir is that: a) water flows through the brush but fish cannot; b) the brush and saplings can be easily replaced from local materials each migration season and; c) the stone "foundation" which holds the brush and saplings can be easily repaired and maintained from stream stones each migration season.

Once a stone and brush weir is built, migrating fish are blocked by the structure and congregate in large numbers directly below it, trying to find a way past it. Once fish numbers are large enough below the obstruction, they can be easily netted, speared or caught in baskets. The problem with a "one weir" system in a streambed is that upmigrating fish quickly disperse downstream from their congregation at the upstream obstacle once people enter the stream to catch them, and only return upstream to the obstacle when the coast is clear.

This paper hypothesizes that the two weirs in China Lake Stream may have been built and operated together to act as a fish trap. If this is true, the trap would have been operated in the following manner. In the spring of each year, as run-off began to subside, saplings were inserted vertically into the stone foundations of the weirs and interlaced horizontally with brush to create a barrier across the stream channel. A narrow chute was left open on the lower weir. As alewives began to migrate up the stream they passed through this chute in the lower weir but their progress was stopped by the impassable upper weir. Over a period of several days, the river channel between the two weirs filled with alewives.

Once the channel was deemed full enough of alewives to make harvesting easy, the lower chute was closed off to prevent alewives from escaping downstream. Then, members of the family group (or groups) waded into the stream channel with baskets and/or nets and easily captured the alewives which could not swim upstream or downstream to escape. Once the number of alewives left in the stream was reduced by the harvest, the barrier at the lower weir was removed to let more alewives upstream into the "trap" until the numbers were again high enough to allow for easy capture. This "refilling of the trap" might have been done several times a day during the height of the run or every few days at the beginning or the end of the run. This method would also allow the group sufficient time to stake and smoke the alewives they had already caught, while letting new migrants to collect in the trap.

This same method and sequence of weirs would be equally effective in capturing adult American eels migrating out of China Lake through China Lake Stream during the fall because it would greatly increase the effectiveness of a spearing or netting/basket harvesting.
A "W" shaped stone fishing weir for alewives, West Branch Sebasticook River, Pittsfield, Maine. First identified by Tim Watts and the author in fall 2002.

---

Tthe China Lake Stream weirs and the West Branch Sebasticook weir in the town of Pittsfield shows a key design difference. The China Lake Stream weirs lack any inverted "V" pointing upstream, and instead are straight-line structures perpendicular to the stream flow. A possible clue is that the river channel at the West Branch Sebasticook site more than 100 feet wide: twice as wide as the China Lake Stream sites. A straight line weir at the West Branch Sebasticook site would be far less effective as a "W" shaped weir at concentrating alewives in a small, confined area in a wide stream channel. On the other hand, the much narrower stream channel at the China Lake Stream site would have made a "V" design unecessary, because the large number of alewives coming up China Lake Stream would have easily filled the narrow channel from bank to bank.

China Lake Stream: A Regional Fishing Center?

The sheer quantity of alewives that could be easily and quickly caught at the weirs at China Lake Stream (hundreds of thousands) would have greatly exceeded the number that could be preserved by a small family group or groups, as would the number of alewives needed to support them. The extensive prehistoric habitation areas along the Sebasticook River near the mouth of China Lake Stream, and the role of the Sebasticook as a highly used east-west travel corridor by Indians as late as the 1700s (hence the construction of Fort Halifax in 1754 to block communication between the Kennebec and Penobscot tribes), suggest the China Lake Stream weir site may have served as a regional fishing center. This is suggested by a document at the Taunton, Mass. Historical Society containing a 1600s account of the regional use of an alewife fishing site at the mouth of the Cohannet River (Mill River) in what is now Taunton, Massachusetts:

"The ancient standers remember that hundreds of Indians would come from Mount Hope and other places every year in April, with great dancings and shoutings to catch fish at Cohannit and set up theyr tents about that place until the season for catching alewives was past and would load their backs with burdens of fish & load ye canoes to carry home for their supply for the rest of the year and a great part of the support of ye natives was from the alewives."

This document further describes the intensive use of alewives by the earliest settlers as fertilizer for corn and the use of children to catch them:

"The first English planters in Taunton found great relief from this sort of fish, both for food & raysing of corne and prized them so highly that they took care that when Goodman Linkon first craved leave to set up a grist mill at that place, a town vote should be passed that the fish should not be stopped. It is well known how much other Towns are advantaged by this sort of fish. Middleboro will not permit any dam for any sort of mills to be made across their river to stop the course of fish nor would they part with the privilege of the fish if any would give them a thousand pounds and wonder at ye neighboring town of Taunton, that suffer themselves to be deprived of so great a privilege ....

"These fish may be catcht by the hands of children in theyr nets while the parents have y'r hands full of work in the busy time of Spring to prepare for planting. Some of Taunton have been forced to buy Indian corn every year since the fish were stopped, who while they fisht, they'r ground used to have plenty of corne for y'r family & some to spare to others. The cry of the poor every year for want of the fish in Taunton is enough to move the bowels of compassion in any man, that hath not an heart of stone."

Caveats:

There are many reasons why this hypothesized two-weir harvesting technique may never have occurred at this site on China Lake Stream. It is possible that the two instream structures found were never built or used as fish weirs; that only one was built as a fish weir; that both structures were fish weirs but were built and operated centuries or millennia apart.

A key factor confounding the truth or falsity of this hypothesis is that there is an extreme paucity of information about any prehistoric fish weirs in streams in the northeast United States. Confirmed prehistoric weir sites and structures in Northeast rivers are extremely rare; and there are almost no eyewitness accounts from European during the Contact Period which describe how these structures were actually used and operated by native people to catch migratory fish such as alewives and eels, which were undoubtedly the target species in China Lake Stream.

With these caveats in mind, a number of evidentiary points suggest this "two weir" fish trap method may have been widely used by prehistoric peoples. The first point is that the general strategy has been widely employed for centuries. The general strategy is that fish are guided into a constricted and confined area in which passive obstructions and barriers are placed to hinder their escape. This is basic principle of the common type of lobster traps and minnow traps sold and used in coastal areas today. This is also the fundamental design principle of "fyke nets" used today to catch juvenile eels in Maine rivers and the many varieties of brush and stick weirs (later equipped with netting) that formed the basis of the enormous weir fisheries for salmon, striped bass, alewives and shad in the tidal estuaries of New England rivers in the 19th century, particularly in the Kennebec, Androscoggin and Penobscot. Baum (1997) reproduces diagrams of many of these weirs, which all use the same design and strategy of that hypothesized for the two adjacent stone weir structures on China Lake Stream, ie. to guide the fish into the "trap" and then hinder their escape with passive barriers.

Future Research Needs:

Due to the advanced age of prehistoric stone weirs, the existence of dam impoundments that flood and obscure them, the cultural overprint of 18th, 19th and 20th century mill dams built at or near the same sites, the effectiveness of stream flow and floods over centuries at destroying or obscuring their remnants, the lack of any focussed effort to locate them, and the skill and intuition necessary to even notice them, it is not surprising that there is virtually no information on the location and types of prehistoric fish weirs in New England and the northeast United States.

The discovery of these two conjectured prehistoric stone weirs at China Lake Stream was only possible due to the very recent (2008) removal of a large hydroelectric dam that had been in place for a century and a directed reconnaissance search by the author on the restored stream reach which specifically focussed on identifying any remnants of a prehistoric weir. This search was driven by the previous historic and biological research which showed that:

a) China Lake Stream once supported large numbers of migratory fish that were actively sought by prehistoric peoples in the area;
b) Extensive archaeological evidence of dense prehistoric habitation along the Kennebec, Sebasticook and China Lake Stream riparian corridors;
c) The 2002 discovery by Tim Watts and the author of a large, prehistoric "W" shaped stone fishing weir on the West Branch Sebasticook River in Pittsfield, Maine and the early 1990s discovery of a wooden stake weir in Sebasticook Lake, Newport, Maine dated to 5,700 B.P.

A key support for the "two weir fish trap" hypothesis presented here would be the identification of multiple sites in the Northeast that fit the general physical description of the China Lake Stream site described herein. If nothing else, this hypothesis has value in alerting investigators to always look for a second weir site just above or below one that has been tentatively identified.

The weir site and adjacent shoreline areas described above would benefit from a professional archaeological survey to determine the full extent of prehistoric use and habitation.

---

References:
Baum, E.T. 1997. Maine Atlantic Salmon: A National Treasure. Atlantic Salmon Unlimited. Hermon, Maine.
Doyle, R.G. 2008. Identification of Lithic Artifacts from Central Maine Coastal Archaeological Sites: A Case Study in Regional Lithic Acquisition Strategies. Flying Passage Press. Gardiner, Maine.
Hanson, J.W. 1852. History of Gardiner and Pittston. William Palmer, Publisher. Gardiner, Maine.
Josselyn, John. 1674. John Josellyn, Colonial Traveler. A Critical Edition of Two Travels to New England. Paul J. Lindholt, editor. University Press of New England. 1988.
Pory, John. 1622. Letter of John Pory to the Earl of Southhampton. In: Three Visitors to Early Plymouth. Reprinted by Plimoth Plantation. Plymouth, Mass.
Watts, D.H. 2004. Maine Historic Engineering Record. Dam Number Five, Cobbosseecontee Stream, Gardiner, Maine. MHER No. 22. Prepared for the Maine Historic Preservation Commission. Augusta, Maine. PDF here.
Watts, D.H. 2003. A Documentary History of the Alewife (Alosa pseudoharengus) in Maine and New England. Friends of the Kennebec Salmon. Augusta, Maine. PDF here.