The Joggins Fossil Cliffs of Nova Scotia, Canada, are one of the world’s most famous fossil sites (Falcon-Lang et al. 2010). Awarded UNESCO World Heritage Status on 7 July 2008, these towering sea-cliffs along the Bay of Fundy have yielded a spectacular array of fossils from the Carboniferous "Coal Age", about 315 million years ago (Calder 1998; Falcon-Lang et al. 2006). Charles Lyell, the well-known Victorian geologist, was an early visitor in 1842 and 1852. He documented upright fossil trees rooted in coal seams and early tetrapods; and later, Charles Darwin alluded to the importance of Joggins in his epoch-making Origin of Species (Calder 2006; Falcon-Lang 2006).

Less well known is that Joggins was one of the first places in North America where coal was mined (Falcon-Lang 2009). French settlers first exploited its coal in the late Seventeenth Century (Franquelin 1686), if not earlier, and the "Coal Cliffs" of Joggins quickly became a prominent feature on maps of North America (Moll 1730). Most early coal mines at Joggins supplied Boston, Massachusetts (Hale 1731), and at times the British and French squabbled over Joggins’ sought-after resources (Knox et al. 1759). However, despite the strategic importance of coal, mining remained sporadic until the mid-Nineteenth Century, when more systematic and larger scale exploitation commenced (Drummond 1918).

Timber pit props and associated structures protrude from the Joggins sea-cliffs at several locations today (Falcon-Lang 2009). These are the remains of formerly underground coal mine workings which have since been exposed by cliff retreat. Dawson (1882) estimated the rate of retreat to be ~0.19 m per year for some parts of the section, based on 40 years of observation. So, several tens of metres of erosion may have occurred since mining commenced. However, although mine workings were first noticed eroding out of the cliff face in the early Twentieth Century (Bell 1913), their age has been much debated and remains uncertain. They may represent late Nineteenth Century mines or they may date from earlier periods of British or French occupation. In this paper, we employ dendrochronological analyses to ascertain the age of the best-exposed coal mine workings.

The Joggins Fossil Cliffs are located on the southeastern shore of Chignecto Bay, the northwestern arm of the Bay of Fundy, Nova Scotia, Canada (Fig. 1; 45°42'N; 64°26'W). Along the 4-km-long cliff section between Lower Cove and McCarron’s Creek, more than 100 coal seams crop out, but only seven attain economic thickness (>1 m thick) and have been mined extensively. From base to top, these are Coal 32 (unnamed), Coal 29A (Fundy, also known as Cumberland or Hardscrabble), Coal 29 (Dirty), Coal 20 (Forty Brine), Coal 14 (Kimberley), Coal 8 (Queen), and Coal 7 (Joggins seams, also known as King or Main), the numbers relating to the systems of Logan and Dawson (Currie 1942; Goudge 1945; Copeland 1959; Davies et al. 2005; Rygel and Shipley 2005; Calder et al. 2006). Coal mine workings, including timber pit props and other archaeological materials, occur on all of these seams, but are especially well exposed on the closely adjacent Fundy and Dirty seams (coals 29A and 29, respectively, as illustrated in Davies et al. 2005), which form the focus of this study.

The Fundy and Dirty seams crop out about 1.7 km south of Lower Cove. Here, the Fundy Seam is a 0.85-m-thick bituminous coal (Hower et al. 2000), whereas the Dirty Seam, as its name suggests, is a 1.5-m-thick clastic-rich (high ash) coal (Calder et al. 2006). Stratigraphically, the coals are positioned at 420 m and 428 m, respectively, above the stratigraphic base of the Joggins Formation and dip at ~21° SSE (Davies et al. 2005). As a result of this inclination, the two coals delineate an outcrop zone extending between 45° 42'14.34" N; 64° 26'31.92" W, where the Fundy seam occurs at the cliff top, and 45° 42'12.36" N; 64° 26'36.90" W, where the Dirty Seam occurs at beach level, a horizontal distance of ~150 m (Fig. 1).

To determine the age of these coal mine workings, we obtained samples for dendrochronological analysis. Sampling occurred during a preliminary investigation in November 2008 (samples MAD Lab 08BDD000) and a follow-up study in the summer of 2009 (samples MAD Lab 09ES000). Permission to sample at this protected UNESCO World Heritage Site was granted through Nova Scotia Museum Heritage Research Permits #A2009NS50 and #A2008NS94.

In total, 73 samples were collected. After microscopic observation of anatomical characteristics, 70 of these samples were determined to be red spruce (Picea rubens). These samples could eventually be compared with an existing regional master chronology for red spruce (MAD Lab archives), which spans the interval 1624 to 2006. Three samples (08BDD01, 09ES17, 09ES18) of white birch (Betula papyrifera) and balsam fir (Abies balsamea) were excluded from the analysis because no long regional master chronology for these taxa is available. A further 22 samples were also excluded from the analysis because they contained less than 30 tree rings. Short tree-ring sequences produced statistical uncertainty in their placement against a master chronology because they might fit into any number of time periods (Grissino-Mayer 2001). The remaining 48 samples qualified for the full analysis, having an average of 94 rings and a range of 33 to 247 rings.

From the Fundy Seam, 35 samples were analyzed and incorporated into a floating chronology from 10 separate openings (1–10). From the Dirty Seam, a further 13 samples were analyzed and incorporated into a floating chronology from 4 openings (11–14). The full dataset is given in Tables 1a and 1b, showing the number of tree-rings in each sample, the estimated cut date for each tree (i.e., the year it was felled, when bark is present), the significance level of correlation between the averaged two radii per sample to the master red-spruce chronology, and whether bark was present in the sample. Where the bark was not identified, estimated cut dates are minimum estimates, as tree-rings could have been removed from the outer part of the tree by decay or milling. However, this problem affected only 10 samples (21%), so we have confidence in nearly all of the cut dates.

Results show a similar age profile for mine workings on both coal seams. The cutting dates for samples with bark on the Fundy Seam are 1849–1875 (mean correlation 0.387 to the master based on 50-year segments), and for the Dirty Seam they are 1856–1875 (mean correlation 0.445 to the master based on 50-year segments). The overall correlation of all 48 samples taken from the cliff to the master chronology is 0.412 (based on 50-year segments) when compared to the master chronology (Fig. 4). The results of this study provide strong dating of the timbers as a group, with 71% of the patterns of the samples being accurately dated at a 99% confidence interval.

Our dating of the mine workings on the Fundy and Dirty seams can be validated against written documentation. Historical records show that the Fundy Seam was worked in the zone close to the cliff edge between 1731 and 1943, albeit sporadically (Hale 1731; Drummond 1918; Goudge 1945; Copeland 1959; Falcon-Lang 2009). Dendrochronological studies presented here illustrate that the timber pit props currently exposed in the sea cliffs were cut between 1857 and 1875, with the exception of one sample dated to 1849. This range of dates therefore allow us to focus on the historical accounts of Nineteenth Century mining activity on the Fundy Seam to test dates independently, and allows the archaeological record to be tied precisely with written documentation. This, in turn, provides additional information about the nature and timing of the coal mine workings preserved on the Fundy and Dirty seams.

In 1827, the General Mining Association (GMA) became "lessees of all the mines and minerals of every description in the province of Nova Scotia Proper" (Martin 1834, p. 105) and opened mines at Pictou and Sydney. However, it wasn’t until 1847 that the GMA opened its first mine at Joggins (Anon 1848; Currie 1942). This mine exploited the Joggins Seam, which crops out ~1 km south of our study site on the Fundy/ Dirty seams. The mine was set up at a cost of £16,000 and managed by Joseph Smith, who also ran the GMA mine at Pictou (Gesner 1849). An 1863 report indicates that the Joggins Seam was initially worked by driving an adit into the coal from its crop on the cliff (Currie 1942) but, soon after, it was worked by a ~30-m-deep shaft sunk ~80 m from the contemporaneous shoreline (Boggs 1871; Rutherford 1871) a little inland of the site of the present-day UNESCO centre (Falcon-Lang et al. 2010). Coal was brought to the surface using a horse-gin, carried down an inclined narrow-gauge railway to a pier on the shore near Bell’s Brook, and loaded onto ships. Production increased from 2400 tons (1851) to 8478 tons (1866) during its initial operation, most of it being exported to Saint John, New Brunswick (Campbell 1871).

In 1857, the Nova Scotia government forced the GMA to give up most of its mineral rights in the province. However, the GMA retained control over the four-square-mile tract at Joggins (Brown 1899; Copeland 1959). In 1865, it expanded operations through the development of a drift mine on the Fundy and Dirty seams (Browne 1868). This was called the "New Mine" and was entered by an adit a few metres above the high water mark on shore (Dawson 1865, 1868). Coal was extracted through a series of slopes driven up-dip from horizontal levels to the crop-line on the cliff top (Hill 1888). Rutherford (1871, p.11), who was inspector of mines from 1865, described this GMA mine as follows: "The hard scrabble seam [=Fundy Seam] is entered by an adit in the cliff, and the coal is taken direct from the mine into a covered shed, out of which it is removed at a lower level, and carried along a wharf to the ships. This is but a temporary arrangement, but it is a very usual one on first opening mines similarly situated."

In 1871, the GMA abandoned the New Mine and sold this part of the Joggins tract to the Joggins Coal Mining Company (JCMC) (Drummond 1918). A report to the Nova Scotia House of Assembly indicates that the JCMC recommenced extraction of Fundy and Dirty seams through a slope mine opened in 1872 after a short lull in activity: "The Joggins Company have opened this colliery on the seam at Hard-scrabble [=Fundy Seam]. The slope has been started a third of a mile from the shore and driven down 270 feet, at which depth levels have been started to the right and left. An inclined tunnel 400 feet long, through the cliff, connects the mine with the wharf. The winding engine has two horizontal 11 inch cylinders, 30 inch stroke, and geared three to one to a drum 8 feet in diameter. The three boilers, 25 feet long and 2 feet 9 inches in diameter are hung from baulks of timber, and are all three fired together" (Drummond 1918, p. 34). This second mine was centred about 500 m east of the GMA mine, but its transport tunnel to the wharf on the shore, and associated levels, passed through the older GMA workings.

The history of the JCMC mine was short-lived. Operations had commenced at the height of a boom in the regional coal industry, triggered by the completion of the Truro-Moncton section of the Intercolonial Railway, in 1872. This railway, which ran through Maccan, less than 10 miles from Joggins, meant that coal could be used by coal-fired trains and easily transported to Halifax, although Saint John remained the main market. By this time, the Fundy Seam had acquired a particular "reputation as a superior coal for house purposes" (Boggs 1874, p. 3) and was widely used in domestic hearths. Thus, when the Great Fire of Saint John (20 June 1877) drastically reduced local demand for domestic coal, the JCMC mine lost its principal market and closed (Brown 1899). From 1877 to 1903, no mines operated on the Fundy and Dirty seams at Joggins, all the work being centred on the Joggins Seam (Goudge 1945).

The cut dates of the timber samples currently exposed at the Joggins Fossil Cliffs (1849–1875) show a good correlation with the written record of mining activity at the site (1865–1877). Our analysis allows identification of two phases of mining. Most of the openings currently exposed are relics of a GMA operation that mined up-dip from beach level from 1865 to 1871. A few other openings adjacent to a prominent tunnel at the base of the cliff represent levels of a later JCMC operation from 1872 to 1877, which was centred 500 m inland, but whose access tunnel to the shore passed through the earlier GMA workings. A slight discrepancy between dendrochronological dates and those gleaned from the written archives may be resolved if the original GMA workings were constructed from timber stockpiled or recycled from the adjacent, but slightly older, Joggins Mine.