I missed another week. I suppose at this point I should just admit that I’m not finding enough time/inspiration and switch to an every-other-Friday planned schedule. I’d still like to post something new each week, I just seem to have trouble doing it this year.
But today I have something good. This week was vacation, and I managed to hike the Keystone Arch Bridges Trail, in Chester, Mass. This short (about 2.5 miles, 4 km) trail has four stone arch bridges, plus the abutments of a fifth destroyed in a 1927 flood, and one of them, shown above, is still in use after 174 years by the successor of the railroad that built it. Today’s post will have a bit of history of railroading, and some information about another nineteenth-century engineer, two actually.
Let’s start with the engineer: George Washington Whistler, father of the famous painter and thus, obviously, the husband of the subject of the well-known painting, Whistler’s Mother (his second wife, Anna McNeill). Unlike many engineers of the time, who learned their craft as assistant to some other engineer, George, born in 1800, attended West Point, where he studied engineering and graduated in 1819, eventually attaining the rank of Major. He served in the Army Corps of Engineers and the Topographical Survey (this was before the Army created a Corps of Topographical Engineers).
It was peacetime and the Army wasn’t doing a lot of construction, so they loaned Whistler and another officer, William Gibbs McNeill, to the Baltimore and Ohio Railroad (the B&O) to survey a route west to the Ohio river. The B&O was the first American railroad, and likely a very interesting place for an engineer to work in those early days. The officers had known each other as cadets, and Whistler eventually married McNeill’s sister. This was the start of a collaboration that spanned the transition from canal-building to the early U.S. railroads. McNeill would go on to do quite a lot more, but he’s not today’s main subject since he wasn’t involved in this bridge.
While working for the B&O in 1828 Whistler was sent to England to study railroad and locomotive design. He brought back information that was used to build the first locally-manufactured steam engines. The Smithsonian also credits him with introducing the steam whistle to American locomotives.
Whistler eventually left the Army, and a short time later (1833) went to work on the Middlesex canal connecting Lowell to Boston. Sometime around then he also worked on the Boston & Lowell Railroad as well as on the Boston & Providence Railroad with McNeill. For the latter, McNeill was apparently the senior member of the team, and thus gets credit for the Canton Viaduct, although Whistler and two other officers assisted.
The Proprietors of the Lowell Canal System had a fairly extensive engineering staff, complete with their own shops for making turbines (and later steam boilers) and they constructed some of the early railroad locomotives for Boston’s new railroads, which were apparently copies of the British Planet locomotive, although engravings show some differences in form. Whistler was the superintendent of that shop from 1834 to 1837. In 1837 Whistler left Lowell to focus on railroad construction.
One of the early railroads of Boston was the Boston & Worcester, which was built between 1832 and 1835. It doesn’t appear that Whistler worked on this one, but if so it was the only one of the set that he missed. By 1836, work was underway to extend the railroad from Worcester to the New York state line as a separate railroad, the Western Railroad. Springfield, on the Connecticut River in western Massachusetts, was reached in 1839, and the complete route to the state line, where it connected with another railroad, was put into service in 1841. The goal was to give New England’s industries an outlet to the growing markets in what is now the central U.S.; back then that was the frontier. In reality, while it did that, it also opened New England to new industries operating in the west, which was probably more significant in the long run.
Now interestingly, one of the abutments of one of the out-of-use bridges bears a stone with the date 1844 carved on it, and I’ve read elsewhere that the bridges weren’t completed until then, although no source was cited. It would seem that they were using temporary trestles when the line opened, and hadn’t finished the bridges at that point. It wasn’t unusual to have a trestle across an in-construction bridge, as those were used to fill the structure with rubble; only the outer faces would be smooth stone. Or perhaps the temporary structures were off to one side, although room to do that was very limited.
Whistler, initially while still working at Lowell, was also “construction engineer” for the Western Railroad, and apparently worked with McNeill on the portion from Worcester to Springfield. McNeill left to work for another railroad at a salary “double…that which had been given to him…by us” per a Western Railroad report (1843, PDF). Whistler is credited as designer of one of the bridges on this section, a more modest structure than the ones I saw. And he’d credited with being the first to use a keystone arch bridge for a railroad. In 1840 he became chief engineer for the final stretch from Springfield across the Berkshires to the state line.
This was the hard part, particularly the long grade up from the Connecticut river. At Springfield the railroad is at 57′ above sea level (17 m). By Chester, 28 miles (45 km) as the track winds it has climbed to 600′ (183 m). In the next 12.6 miles (20.2 km) it climbs to 1,459′ (445 m) before reaching the summit at Washington (named for the president, not the engineer). The route is a hard one, with a portion built high up the sides of a narrow river gorge. One of the bridges rises 67′ (20 m) above the river. This grade has long been a problem for heavy freights.
The necessary grade isn’t really steep by modern standards, at 1.67% (I’ve also seen 1.65% noted). The B&O had one grade of 2.2%, and that went on to become something of a maximum specification for later mainline railroads (per Trains magazine, Sep 2011 issue). But up to then, railroads had mostly followed relatively level paths. And to make matters worse, it was (and is) a very curvy route, and curves make for more friction, requiring more power to pull the train. Apparently some of the early plans involved using horses, but Whistler managed to do it with steam locomotives the whole distance. Getting over the Berkshires by steam in 1841 was an amazing achievement.
Whistler seems to have been something of a perfectionist. When other railroads built cheap to start, and then upgraded when the money came in, often relocating whole lines in the process, he surveyed a route that’s survived largely unchanged down to the present. And where other railroads started with wooden trestles, he built massive stone bridges wide enough for two tracks to begin with (he did compromise to the extent of only laying one track).
That may not be perfectionism so much as practicality though. These bridges were in a remote location subject to heavy spring floods. A washout would be difficult to repair, not to mention expensive. Stone would also be less likely to catch fire. Early wooden bridges often had tenders assigned to them, men who could put out fires sparked by burning embers from the locomotives. That probably wasn’t practical here. And as for the route: it follows a river, and without major engineering, it’s not going to get any straighter. There actually was some of that: the other bridges are unused today because of a straightening project performed in 1911 that reduced the number of river crossings and made the curves wider on this section.
But back to Whistler’s achievements: In the 1830’s, the typical locomotive was a derivative of Stephenson’s Planet that weighed 12 tons (11 metric tonnes). For his steep grade, Whistler commissioned several 19.5 ton locomotives (18 tonnes) base weight, or 22.8 tons (20.7 tonnes) loaded with fuel and water. These were (per this book and the report cited below) Ross Winans’ Crab design with eight small connected driving wheels, a gear drive and a vertical boiler. The first, named Maryland, was purchased in 1841 after it had been in trials with the B&O. Although the number has been given as three in some sources, there actually appear to have been seven of them purchased as a single order.
The gear-driven eight-connected locomotive appears not to have been a very successful design, nor were the ugly devices well-liked by the railroad. One director of the corporation referred to their purchase as “rash investments in engines of new patterns”, and this may have been what led Whistler to leave the Western Railroad in 1842. Little documentation has survived about them. In any case, they were more than able to haul trains up the grade, even if they were later replaced by other designs that were more successful. Several tests carried out after Whistler left proved that they could out-haul other contemporary locomotives, and use less fuel doing it. Whistler’s choice was a success.
Given the paucity of information on these locomotives, it’s worth saying a bit more about them. The 1843 report noted above is an account of a stockholder-initiated investigation into several matters of financial concern as the new Western Railroad moved from construction to operation and began to worry more about being competitive and generating a dividend. Among these was the choice of locomotive. In 1838 a committee led by McNeill had recommended the purchase of “six ten-ton and two fourteen-ton” locomotives, with Whistler to be responsible (as “chief engineer”, so maybe he held that position before 1840) to “make the drawings and superintend the work”. For the Berkshire portion of the line, Whistler selected three Lowell engines of the existing design (likely Planet-derived designs) for passenger use, seven from Winans for “merchandise” (based on the Crab design) and two small Gillingham & Winans locomotives (which were used over a connecting line owned by another railroad that was built too light to support the Western’s usual locomotives). It’s also worth noting that the term “Crab” was a derisive nickname applied by those who didn’t like the engines; they were otherwise known as “the Winans”. But the name stuck, and was applied to later designs of similar locomotives as well.
The Winans merchandise locomotives were selected after Whistler was dispatched to Baltimore by a selection committee to investigate claims of a new and powerful engine design, which was in use on the B&O. After positive comments from that railroad, Whistler viewed a trial and after he returned to Massachusetts the committee dispatched two members to view additional trails and contract for seven locomotives if they proved suitable. They were, and the purchase was made despite the locomotives costing more than 20% above competing designs, and 50% above what one competitor apparently offered. While it appears that a selection committee made the choice, not Whistler, as chief engineer it was clearly his responsibility to approve the choice. And the selection of a new and untried technology seems to fit the character of a man who was the first to use a stone arch bridge for a train, and whose constructions reflected an emphasis on technical suitability over cost. I think it’s fair to say that he chose them.
As for the locomotives, all seven were delivered and put to use, though not without problems. At least three of the seven were built under contract to Winans by Baldwin and Vale, which is probably the origin of the confusion over the number purchased. The Maryland was delivered in January of 1842, followed by the Michigan, Ohio, Illinois, Arkansas, Indiana and Missouri through the rest of the year. The high cost of purchase, and the decision of the B&O not to purchase any after their own trials, plus some initial mechanical issues, created something of a controversy for the railroad. This was fanned by an unhappy loser, William Norris.
Norris was, however, in some kind of financial trouble and probably needed the sale badly, and he was found to have exaggerated the results of some B&O testing, so his comments are hard to take at face value. His more conventional locomotives, likely his 2-2-0 design, if built, might have met the railroad’s needs at a lower cost. Or not. He had previously demonstrated that the design could pull loads up a grade, but they were never tested on the actual Berkshire route.
In the Winans locomotives’ first year of service, traffic was even higher than projected, and the engines were often overloaded pulling trains beyond their design specifications. This led to mechanical failures (in some cases due to undersized mechanical parts that were then redesigned). Inexperienced crews also damaged the boilers of a couple of them. Despite this, the committee reviewing the financials found that they cost less to run, per mile, than the Lowell engines had in their first year, and that the overall freight operations had been more profitable than those of competing railroads, despite the severe grades. This was probably helped by the fact that it turned out that more freight was coming from Albany than going to it (i.e., more was going down the steep hill than up it), but still the much-disliked Crabs appear to have been a success.
Maybe they weren’t the best possible choice at the time, maybe they were. Even hindsight can’t answer that question. Longer-term the design had limitations, particularly in terms of low speed (top speed was just 12 mph, 19 km/h; acceptable in 1841 but not for long). Winans had problems with his later models not working as promised. In any case, the industry turned away from them and to the precursors of what we now think of as steam locomotives: a smaller number of large driving wheels, with separate wheels to help spread the weight of a larger boiler and firebox. In short, the design Norris had proposed.
In later years gear-driven locomotives would only be used on very steep railroads, mostly logging lines rather than general freight mainlines. Later locomotives purchased by the Western Railroad appear to have been more conventional, although the steep grade still required specialized locomotives. An entire class of steam locomotive, the Berkshire, was designed (c. 1925) to deal with the problem steep mountain railroad grades posed, and gained its class name from adoption for use on this route.
Today, Whistler’s remaining in-use bridge still carries two tracks, and over it run trains he couldn’t even begin to imagine. A typical modern freight will be pulled by two 4,000 hp (3 kw) locomotives weighing in at 196 tons (178 tonnes) each, and they have to work at it. The sound of a set pulling a freight train uphill is quite impressive. It may not be as well-known as the Canton Viaduct, and it doesn’t carry high-speed rail, but a substantial portion of the freight to and from southern New England still passes over Whistler’s bridge, as it has for 174 years (or 171 years, take your pick). That’s quite a legacy.
Whistler wasn’t done making his mark on railroading either. In 1842 he went to Russia to help design the Moscow to Saint Petersburg Railway (Winans was involved in this, too, and sold the railroad a number of his Crabs). There he apparently selected the five-foot gauge for the track (it was popular in the U.S. at the time). That gauge is still used by Russian and former Soviet-bloc railways today. He completed that project, and did several other engineering works in Russia before dying of a heart attack in 1849 after being weakened during a cholera epidemic. He never returned to work on more American railroads. McNeill did, and had a long career. Without a doubt he’s a more significant engineer in the history of this country. But Whistler and his achievements should not be overlooked in the shadow of his friend.
Back to the bridge: the stone arch bridge is an ancient design. The Romans used them widely, as well as using true arches (arches with keystones) in other structures. They are amazingly solid and stable, and added weight only makes them stronger. Many of those Roman bridges remain in use today. So it’s not really surprising that Whistler’s bridges are still with us. And yet, locomotives today weigh ten times (or more) what those heavy Crab locomotives weighed. Freight cars are similarly heavier. At a conservative guess, more than half a million trains have crossed the bridge at the top of this post since it was constructed. Whistler built for the ages. And that, I think, is something any engineer would want to be remembered for.
As a photo, I like this image because it’s a beautiful stone structure in a lovely location. Fall isn’t fully advanced yet, but the yellowing leaves work well with the late afternoon sun. Photographing it back-lit wasn’t ideal, but the only way to get to the downstream side without trespassing involves wading under the bridge, and I didn’t feel like making the long drive home sopping wet. And the strong afternoon light passing under the bridge actually gives this image some character it wouldn’t have if shot from the bright side. I did have a problem with some lens flare on the far right, but cropping out the trees on that side (along with some foreground rocks) eliminated almost all of that.
This image is, as usual, an HDR made from five exposures. The bracket was centered at 1/200th-second exposure, at ISO 400, f/11, with the lens fairly wide at 32 mm. I actually took a number of brackets at different settings trying to find one that kept the sky from blowing out while still preserving dark details due to the high contrast (I probably should have used 2 EV steps rather than 1 EV for my bracket). I did not use the circular polarizing filter (it was too dark, and shooting into the sun isn’t likely to yield much benefit from one anyway). As noted I cropped it to fit. Adjustments were fairly minor. I had to back off exposure a bit, and applied a neutral-density filter to the top portion. Back in C1 I applied the usual sharpening before export.