ECONOMICS OF ROAD CONSTRUCTION BY HALBERT POWERS GILLETTE SECOND EDITION SECOND THOUSAND NEW YORK THE ENGINEERING NEWS PUBLISHING COMPANY 1908 COPYRIGHT, 1901. 1906, BY THE ENGINEERING NEWS PUBLISHING Co. CONTENTS Chap. Page. I. Historical Review 1 II. Earth Roads and Earthwork 3 Profile of Cross-Section of Road 3 Longitudinal Profile 6 Gutters and Drains 6 Embankments 7 Cost of Earthwork 8 Surfacing 11 Traction and Tractive Power 13 Location 15 III. Gravel Roads 16 IV. Macadam Roads 19 What Holds Macadam Together 19 Quality of Stone 23 Relative Wearing Powers of Stone 24 Quarrying T 25 Drilling 26 Dynamite -. 27 Crushing 27 . Hauling 29 Spreading 30 Rolling 31 Sprinkling . . 32 Quantity and Cost 32 V. Telford Roads 38 VI. Repairs and Maintenance 40 Continuous vs. Intermittent System 40 Sandstone Macadam 41 VII. Suggested Improvements in Existing Road Specifica- tions 44. Kind and Sizes of Broken Stone 44 Depth of Pavement 45 Final Surfacing 46 Material for Embankments 47 Thickness and Width of Pavement 47 VIII. Summary and Conclusion 48 ILLUSTRATIONS Fig. Page, 1. Standard Cross-Section for Macadam Roads, Massa- chusetts Highway Commission, - 4 2. Cross-Section of Road with Gentle Curves and Shallow Ditches, - - 4 3. Macadam Road with Deep Ditches, - 6 4. Macadam Road with Shallow Ditches, - - 7 5. Deep Ditch and Culvert Partially Filled with Dirt Wash- ings, - 10 6. Macadam Road on Steep Grade with Shallow Paved Gutters, 11 7. Typical Cross-Section of Macadam Road, as Recommend- ed by the Author, - - 25 8. Macadam Road with Shallow Ditches, after Winter's Use, 29 9. Another Example of Shallow Ditches, - - - - 36 ECONOMICS OP ROAD CONSTRUCTION. Chapter I. HISTORICAL REVIEW. The roads of the Romans were built at enormous cost, probably not less than $50,000 a mile. They were constructed of stone laid in mortar, the courses aggregating about 36 ins. in thickness. No regard was paid to topography. Hills were climbed or excavated, even where by a slight deflection they could have been avoided, and drainage ditches seem to have been entirely omitted. It was not until 1764 that Tresaguet, a French engineer, began to con- struct roads upon scientific principles by using small broken stone placed upon a well-shaped bed of larger fragments set on edge. The same type of road was later introduced into England by Tel- ford .about 1824, and roads so constructed are generally known by his name. About the same time, or possibly a little earlier, a genius ap- peared in the person of Macadam, who discovered that the foun- dation course of large stone set on edge was unnecessary ; and that, provided the soil was well drained, a bed of loose broken stone some 6 to 12 ins. in thickness would become sufficiently bound together under traffic to uphold the heaviest of wagon loads. No further advance in the science of road construction occurred until the invention of the stone-crusher and the intro- duction of the steam road roller, both within recent years. It may be safely said that half the controversies between engin- eers over the subject of road construction are due to the fact that one class of engineers still adheres to the rules of Macadam with- out recognizing the changes wrought by the steam roller and the crusher. The broken stone road of to-day is quite a different structure from the type of road built by Macadam, who used hand broken stone that was practically uniform in size, laid in the road without the addition of a binder of stone dust or sand and left to be compacted by passing wheels. What was the result? The wheels cut ruts in the loose stone until the soil worked up from below, while the action of the wheels powdered and broke some of 2 EARTH ROADS the stone until the voids were filled, and the mass became packed, weighing 90% to 95% as much as solid stone. In this process it took 18 ins. of loose broken stone to make 12 ins. of macadam surface, or 1.5 cu. yds. of loose stone to 1 cu. yd. of macadam. To this day it is therefore stated in every text-book known to the writer, that the steam roller will compress loose stone one-third, or that 6 ins. will roll to 4 ins. ; which is one of the errors that seems never to be contradicted. Rolled as roads now are with a steam roller, no such compres- sion as this is possible ; although in cases where the stone is placed upon loose unrolled earth sub-grade some stone is driven into the earth and lost, which has led many engineers to believe that the roller had compressed the stone 33%, or even more. This is men- tioned as but one of the errors commonly accepted as truth, and one that is in a measure accountable for too high estimates of the amount of broken stone required on the one hand, and too low an estimate of screenings on the other hand. The true shrinkage will be given later. With the introduction of the rock crusher came a large amount of stone dust or screenings (%-in. diameter and less). These were at first rejected as being worthless, in fact detrimental to good road construction, but some one with more brains than book-learn- ing tried them, and found that they made the road bind more quickly and gave better results than were attained by following Macadam's specifications. The use of a binder in conjunction with the steam roller then made it possible to build a good road in a few days, where formerly it had taken months ; and it led to the building of "thin roads," of 6 ins. and even 4 ins. depth of road metal. Such a radical departure from precedent was and still is ridiculed by some engineers, but common sense and economy seem now to be winning the day. We have seen the thickness of broken stone roads reduced from the 36 ins. of the Romans to the 18 ins. of Telford, to the 12 and even the 6 ins. of Macadam, to the 6 and in some cases the 4 ins. of to-day. The recent achievements in economic road construction are due entirely to three factors: (1) Proper drainage and rolling of the earth foundation ; (2) The use of machine broken and screened stone with the screenings for a binder; (3) Thorough consolida- tion with a steam roller; and it is safe to say that an economic road cannot be built unless all of these factors enter into its con- struction. There is yet another factor that up to the present has EARTH ROADS 3 9 been ignored by engineers ; namely, the use of machines for grad- ing. Contractors have been, and are, well aware of the great econ- omy attending the use of drag and wheel scrapers, of "road ma-; chines ' ' and Shuart graders ; but the cross-section of roads de- signed by engineers is usually such that the use of these machines is practically impossible. A reference to the reports of the Massa- chusetts Highway Commission will disclose the fact that about 5,000 to 6,000 cu. yds. of earth excavation are made per mile of road, at an average cost of about 33 cts, per cu. yd. ; wages being $1.50 for labor and $4.00 for team and driver for nine hours. It can be positively stated that both the amount of excavation and its cost can be greatly reduced, and probably 50% saved, sim- ply by a change in the cross-section of the road, permitting the use of scrapers, and by a change in specifications, permitting some vegetable matter in the embankment, and the requirement of less labor in slicking, or "sandpapering," the slopes of embankments, and omission of sprinkling and rolling during construction. This statement will become more clear upon perusal of the next chapter. Chapter II. EARTH ROADS AND EARTHWORK. The cheapest in first cost and consequently the most common form of road is one made entirely of earth, properly crowned and rolled, either by wheels, or by rollers ; and in any case, whatever may be the paving material, the cross-section should be designed along the lines now to be described, if true economy in construc- tion is desired. PROFILE OF CROSS-SECTION OF ROAD. It is too prev- alent a practice to design a uniform cross-section for a road, regardless of the soil of which it is made ; regardless of the climatic conditions, and the drainage area that the ditches must serve ; regardless of the inclination or slope of ditches, and regard- less of side-hill or other excavation. As an introduction to the consideration of these factors present practice in road construc- tion will be discussed, as exemplified by the standard cross-section used and recommended by the Massachusetts Highway Commis- sion, 4 EARTH ROADS The ditches are commonly made 3 ft. deep, 1 ft. wide at the bot- tom, with side slopes of 2 to 1, as shown by Fig-. 1.* This great depth is given to the ditches, we are told, for the pur- pose of thoroughly draining the soil under the road, so that the frost will not heave or destroy the road surface, a theory that the writer has not found to be sustained by experience. Even granting that a shallower ditch would leave more moisture under the road that would freeze, it becomes a very important question as to the effect of such freezing, and the writer has never seen any deleter- Mocadam FIG. 1. STANDARD CROSS-SECTION FOR MACADAM ROADS, MAS- SACHUSETTS HIGHWAY COMMISSION. ious results in ordinary soil where the road bed had been drained by a ditch, whose bottom was 18 ins. below the crown of the road. As to theory, it seems probable that any expansion of the moisture in the voids of the earth upon freezing would be taken up by the 12-in. layer of dry earth between the frost and the macadam ; and should it not be thus taken up, it would merely raise the whole surface of the road uniformly a fraction of an inch. It may be argued that "faulting" may be caused by the lateral pressure of frozen earth ; if so, the line of weakness is in the ditch where the bulging will occur and not under the macadam ; and it is erroneous to assume that expansion will take place along the axis of the H FIG. 2. CROSS-SECTION OF ROAD WITH GENTLE CURVES AND SHALLOW DITCHES. road, for ice melts under pressure and flows toward the point of least resistance. Since the writer formulated this theory to account for the fact that freezing does not injure macadam, even where shallow ditches are used, he has received conclusive proof of its truth. Prof. Daniel B. Luten's recent experiments upon the expansion of gravel, sand, and loam, saturated with water, and frozen in an iron tube, 36 ins. long, filled to within 3 ins. of the top, showed an expansion of % i n - with gravel, % in. with loam, and an imper- *The half-tone illustrations presented further on, with explanations and comments beneath the titles, illustrate the two types of cross-section, as shown by completed macadam roads. CROSS-SECTIONS 5 9 ' ' ceptible amount with sand. The rest of the expansion was taken up by forcing water up through the earth, and forming a plug of ice at the top of the tube ; a 1-in. ice plug on the gravel, a y^-in. plug on the loam, and a i^-in. plug on the sand. Under a road, when freezing begins, the expansion, in a similar manner, will force a small amount of water up into the voids of the dry layer beneath the macadam, out into the ditches, and down into the earth below, thus causing no perceptible heaving of the macadam. It must be remembered that a soil is seldom sat- urated with water as in these experiments and therefore little or no expansion of the soil itself will take place under ordinary con- ditions. Since it is evident that the deep ditches are usually un- necessary, it follows that broad, shallow ditches are better, not only because less earth need be moved to form them, but because by giving the ditch a broad (24-in.) bottom, drag scrapers can be used in moving the earth. This leads to another criticism of the common design of road cross-section as made by engineers who seem not to have considered the great economy attending the use of drag and wheel scrapers and road machines. Deep, narrow- bottomed ditches, and the "shoulders" upon each side of the ma- cadam, shown in Fig. 1, render the use of such machines generally impracticable. The wheel scraper is the greatest labor-saving device for moving earth short distances that was ever invented, and road machines and leveling scrapers stand second only in point of merit. It can be safely said that where these machines can be used the cost of earthwork will be not over two-thirds what it is where pick and shovel work is necessary. Wide, shallow ditches are favorable to the use of drag and wheel scrapers and all that remains to render the use of leveling machines possible is a gentle curving cross-section profile of the finished earth surface, as shown in Fig. 2. The cross-section of the ditch shown in Fig. 2 will give sufficient earth to build the embankment forming the earth road. Here again the engineer is very likely to err, particularly if his training has been in railroad work. He will balance his cuts and fills on the longitudinal profile of the road, and entirely forget the large amount of earth that will come from the ditches. It is therefore a good rule to balance in the usual way the cuts and fills on the profile of the center, and afterward raise the grade about 3 to 6 ins. all along the road, to make room for the earth from the ditches. While Fig. 2 may be taken as a standard cross- section, it should be modified to suit local conditions. In very wet, soggy 6 EARTH ROADS soil, or where the drainage area of the ditches is great, the ditches must of course be made larger; while in very sandy soil little ditching is required, for a certain amount of moisture in sand is desirable to hold it together. LONGITUDINAL PROFILE. Another common error made by railroad and canal engineers is to design level or nearly level grades for long stretches of the road. In road work, however, the minimum grade should be y%% to secure longitudinal drainage of FIG. 3. MACADAM ROAD WITH DEEP DITCHES (Showing a New York State macadam road of "standard cross-section" with ditches of extreme narrowness at the bottom and of unnecessary depth. Owing to not balancing cuts and fills, material has been wasted out- side of ditch line and in front of fence lines.) any incipient ruts that may form in the road surface ; and this is especially necessary in cuts where the water seeps through and runs down the slopes of the cut, frequently washing ditches full of earth where they have only a slight grade. This leads us to the consideration of cuts. Deep cuts form the most expensive portions of a road, both in first cost and maintenance. To reduce the first cost of earthwork in cuts, the surface ditches should be narrowed until they are merely gutters. GUTTERS AND DRAINS. Gutters must be paved with cob- blestone on grades of 2% or more. Underneath the gutter, tile drains must be placed to carry the water fed by the ditches above GUTTERS AND EMBANKMENTS ? 9 the cut. It should be observed that after every heavy fall of snow the mouths of all such drains and all pipe culverts must be cleaned out, otherwise water will dam back and under pressure will either overflow the road or follow along the pipe and undermine it, causing- bad settlements. Regarding the width of the road, it should never be less than 15 ft. between ditches, and preferably 22 ft. or more. The surface should not have a side slope greater than %-in. in 12 ins., other- FIG. 4. MACADAM ROAD WITH SHALLOW DITCHES. (Showing- macadam (6 ins. x 16 ft.) road without ditches; earth shoulders 7 ft. wide; crown of road 15 ins. above bottom of earth gutter.) wise traffic will follow the ridge in the center, and more quickly wear it out. EMBANKMENTS. We come now to the formation of em- bankments. The scissors and paste-pot engineers usually specify that all material of vegetable nature must be carefully excluded from the embankments. This is unnecessary and adds immensely to the cost where the road passes through meadow land. It is of course desirable to remove brush, large roots, and high weeds ; but the writer has built embankments in the densely timbered and brushy country of western Washington, where no efforts were made to exclude small twigs and sod, and after eight years em- bankments showed no undesirable settlement. An embankment 8 EARTH ROADS built with wheel or drag scrapers, wagons and horses, needs no rolling until its surface is shaped and ready for completion; indeed the "sectional iron roller," often specified, will produce no greater consolidation than will the wheels and hoofs. An embank- ment need not be made in horizontal or concave layers, as fre- quently required by the engineer. All such "sandpapering" is the work of a theorist, which experience proves unnecessary. In passing, the writer will call attention to a common error, namely, that embankments shrink some 10$ after construction. Embankments made with pick and shovel and wheelbarrows do settle under the puddling action of rain, but in embankments made by teams the shrinkage takes place almost entirely during the construction, under the pounding of hoofs and wheels; and even very high embankments so constructed do not ordinarily settle over 2% afterwards. The error has probably arisen from consideration of investigations of Morris and others, showing a shrinkage of 10% when earth is taken from "cut" and put into "fill." It is true that shrinkage takes place, but the shrinkage occurs during the process of construction, not afterward.* COST OF EARTHWORK. The data of Morris, as copied and erroneously modified by Trautwine, are so commonly quoted as being exact enough for practical purposes, that the writer wishes here to state that not one of Trautwine 's tables of costs comes within 25% to 50% of the truth for short hauls, either with carts, scrapers, or wheelbarrows. Trautwine unfortunately underesti- mates the cost of all earthwork except for long hauls ; and there his tables are useless, since two-horse wagons, which he does not mention, are cheaper than carts. The writer has published three handbooks! for engineers and contractors, giving all his data and conclusions in full ; but for the present 'purpose it will suffice first to point out some of the causes of Trautwine 's errors; and second to give the writer's own rules for estimating costs. Trautwine errs, (1) in assuming that a wheel scraper or drag scraper holds the amount of earth that its catalogue size would indicate, because it will usually not go out full ; and even if it does it is full of loose-ploughed earth that will shrink 25% when rolled and packed by hoofs and wheels; he errs (2) in too low esti- mates of time lost each trip in loading and unloading; he errs (3) in the speed of horses going at a walk ; they travel not 150 ft. a minute as stated by him, but 220 ft., which in a measure counter- *See Engineering- News, Nov. 15 and 22, Dec. 13 and 20, 1900. fEarthwork and Its Cost, $2.00; Rock Excavation, $3.00; Handbook of Cost Data, $4.00. COST OP EARTHWORK 9 ' balances some of his other errors in short hauls ; he errs (4) in assuming that one driver can look after three or more carts. In Trautwine's table of costs of moving earth with drag scrapers it is stated that with leads of 40 ft. one drag scraper will move 220 cu. yds. in 10 hours, which is absolutely impossible. The writer has never seen more than 70 cu. yds. moved with one scraper in a day and then only in the easiest of soil and with shortest of hauls. It is high time that engineers should cease quoting such erroneous tables of costs tables that have caused many contractors great financial loss. The following are the writer's rules for ascertaining cost of moving earth, not including superintendence or contractor's profits, wages of laborers being 15 cts. per hour, of team and driver, 35 cts. per hour : The length of lead is measured in feet from the center of grav- ity of the cut to the center of gravity of the embankment, no allowance being made for turning around at either end, as this is included in the "fixed cost." The fixed cost includes cost of loosening, loading, dumping, spreading and lost time of men and teams, but does not include superintendence or contractor's profits and use of tools, for which at least 25% should be added. It is absurd to talk of 10% or 15% profit in earthwork, especially in a new and unknown soil. A team can travel at a walk 25 miles in 10 hours, or at a speed of 220 ft. a minute. A laborer will load with a shovel about 15 cu. yds. (place measurement) of ploughed earth in 10 hours. A plough, team and driver with one man holding the plough, will loosen 400 cu. yds. of ordinary earth in a day ; but if the hard earth crust of an old road must be loosened, it will take two teams, one man riding the beam and one plough holder, to loosen about 150 cu. yds. a day. The fixed cost of moving ordinary earth with wagons is there- fore as follows : Cts. per cu. yd. Ploughing 1 Loading wagon 10 Spreading . . , 2 Time lost waiting to load and unload 5 Total fixed cost of earth in wagon. . 18 Ill EARTH ROADS To this 18 cts. add 4 to 6 cts. if very tough clay or old. com- pacted crust must be ploughed. The cost of hauling is as follows: In soft, loose soils where wagons must be pulled up steep (10'/$ ) embankments, 2-3 cu. yd., place measure, is an average load, while over fair earth roads with steep pull 1 cu. yd. forms a load; on hard, level earth roads, l 1 /^ cu. yds. is a load. Rule To the fixed cost of 18 to 22 cts. per cu. yd. in the wagon FIG. 5. DEEP DITCH AND CULVERT PARTIALLY FILLED WITH DIRT WASHINGS (Showing- 12-in. culvert on New York State road after nature has begun to obliterate unsightly ditches. If it is argued that so large a ditch is neces- sary to carry the surface water, why are culverts made of less than one- eighth the carrying capacity of the ditch?) add 0.8 cts. per 100 ft. of lead in soft roads, 0.55 cts. per cu. yd. per 100 ft. over fair roads, and 0.4 cts. per cu. yd. per 100 ft. over good earth roads. Drag scrapers hold on an average only 0.1 cu. yd. and the fixed cost is about 10 cts. per cu. yd. of ordinary earth for leads of 100 ft. or less, to which add 6 cts. per cu. yd. per 100 ft. additional lead. No. 1 wheel scrapers average 0.2 cu. yd. per load and the fixed cost of earth in the wheeler is 10 cts. per cu. yd. for 100 ft. lead or less, to which add 3 cts. per cu. yd. for each additional 100 ft. lead. No. 2 wheel scrapers average 0.25 cu. yd. per load, so that the SURFACING 11 cost of hauling is 2 1 / 1 > cts. per 100 ft. ; but where a snatch team is necessary in loading, 3 cts. per cu. yd. is added to the fixed cost. To these figures for cost by scrapers or wheelers add 5 cts. if tough clay or hard road crust must be excavated. A study of the above will disclose the fact that for short hauls no method compares in economy with the drag and wheel scrap- FIG. 6. MACADAM ROAD ON STEEP GRADE, WITH SHALLOW- PAVED GUTTERS. (Showing macadam (6 ins. x 12 ft.) road with broken stone gutters (4 ins. x 5 ft.) on 5.3 per cent, grade up which a team hauls a gross load of 12,000 lt>s., exerting a tractive force of not less than 800 Ibs.) ers ; indeed, earth can be moved for about two-thirds what it will cost by wagons, up to 200 ft. lead. SURFACING. The surface of the sub-grade should be leveled either with a Shuart grader or one of the common road machines. In cuts the ploughing should extend below the sub-grade to loosen up the earth so that the blade of the machine will more easily fill up the depressions. A grader will readily level 25,000 sq. yds. in a day at a cost of 0.02 cts. per sq. yd. If the earth road thus pre- pared is to be macadamized, it should be thoroughly rolled with a steam roller so that no stone will be afterward pushed down into the earth. The contractor cannot be too painstaking in this respect, for his pocket-book's sake, if he is to be paid for the stone by the cubic yard in place. 12 EARTH ROADS In loose sand the roller cannot be used at all without first cover- ing the sand with a layer of clay or loam, or thoroughly soaking the sand with water, the latter practice often being very expens- ive. An inch of fine dust spread over the sand and washed in to fill the voids will be found effective and usually less expensive than the use of cotton cheese-cloth recommended by Massachusetts engineers. Half -inch to l^-in. screenings from the crusher may also be used to advantage; for by spreading a layer one stone thick and rolling it, no more stone will be lost in the sub-grade. A thin coating of straw has also been used to hold up the macadam on a sandy soil, and other expedients will suggest themselves to the engineer not hide-bound by precedent. We have thus far considered the itemized cost of moving earth and forming the surface. The following two examples, taken from the writer's timebooks, show the total cost of grading a mile of road under varying conditions. The first case shows cost of work done with that of Fig. 2, designed by the writer. Case I. Cost of grading and surfacing one mile of road in or- dinary earth, gravel and clay: 90 days, team and driver on scrapers, at. . . $3.50 $315 60 days, labor on scrapers 1.50 90 10 days, foreman 2.50 25 Total, 3,000 cu. yds. excavated at 14 1-3 cts. $430 In Case I. all hauls were short, none being over 500 ft., and the average not over 200 ft.; no "sandpapering" of slopes was re- quired, but a good, substantial, workmanlike job throughout. Case II. In this case deep ditches were dug and carefully trimmed, the cross-section being like Fig. I. As the time-sheet shows, there was very little teaming, but nearly all hand-work. Cost of grading and surfacing one mile : 60 days, team and driver, on plough and wagon, at $3.50 $210 460 days' labor . 1.50 690 40 clavs, foreman. . . 2.50 100 Total, 4,500 cu. yds. excavated at 22 1-5 cts. #1,000 There was less hauling in Case II. than in Case I., but most of the earth was either wasted on the sides of the ditches or thrown directly into the road, and careful trimming was required, while TRACTION 13 the shoulders and general form of the cross-section made it impos- sible to use graders. These two cases strikingly and accurately illustrate the differ- ence between good and bad engineering design. In Case I., using wheel scrapers, a good road was built with 30% less excavation and at 35% less cost per cubic yard than in Case II., proving the writer's contention that road design, so far as earthwork is con- cerned, may be greatly improved both in Massachusetts and New York State. The cost of earthwork and of earth roads has been discussed, not to advocate a road surfaced with earth, but because an earth road must be made upon which to build the macadam or Telford pavement. An unpaved earth road will, it is true, remain in good condition when travelled by wide-tired wagons until the fall rains begin to soak its surface, when nothing can prevent rutting first, and disintegration by frost afterward. It is frequently stated that an earth road is as good as any other kind of road if kept well drained and free from ruts, which re- minds the writer of an old saying that a certain bronze eagle in Salt Lake City "flies down to get a drink every time it hears the town clock strike." The statements are true in both cases, but the conditions are equally impossible of fulfillment. It is not a fact that an earth road is as good as any other type, even if kept free from ruts, as the accompanying table of the number of pounds tractive force necessary to pull a ton over dif- ferent surfaces clearly shows : Lbs. Street car tramway 20 Asphalt 25 Stone or wood block pavement (good) 30 Macadam or plank road (good) 35 Macadam or plank road (poor) 50 Gravel, good hard road 75 Clay, good hard road 100 Earth, loose 300 It is evident that even with the best earth road only half as great a load can be hauled as on a good macadam road, which is in itself sufficient to condemn an earth road for any but a poverty- stricken community. TRACTION AND TRACTIVE POWER. This leads naturally to a brief consideration of traction and tractive power. All au- 14 EARTH ROADS thorities agree that a horse cannot exert a continuous pull of more than 100 to 150 Ibs. for eight to ten hours, going at a speed of 2 l / 2 miles an hour, and authorities state that for a short period of time a horse may exert double his average tractive force, or about 250 Ibs. The writer has found that a horse may exert 500 Ibs. tractive force without injury for at least two hours out of ten. He has used a team for raising a 2,000-lb. pile hammer, where the power was multiplied three times by pulleys, and the team has worked about three hours a day, actually lifting a weight equivalent to about 666 Ibs. dead lift, which, with friction, was not less than 1,000 Ibs. tractive force exerted by the team. Tests made by the U. S. Agricultural Department showed that a pair of small mules exerted a continuous pull on a trachometer of 1,000 Ibs. while hauling a wagon up a steep hill. (See Fig. 6 for a view of finished macadam road with 5.3% grade.) It may be safely said, therefore, that a team can exert four times as much tractive energy going up a short hill as its average pull upon the level. Each added l ( / f of grade is equivalent to an added resistance to traction of about 20 Ibs. per ton; therefore, if the average load of a wagon is 8,500 Ibs., and the weight of the wagon is 1,500 Ibs., the tractive force necessary to pull it over a good macadam road on the level would be 175 Ibs., or, over a poor macadam road, 250 Ibs., which would still leave about 750 rbs. available tractive force that a team could exert going up a short hill, such a hill having a grade of 7 1 /i> < /^ It is probable that on a smooth macadam road, and on a grade as steep as T 1 /^, the horses might slip and fall, so that some allowance should be made for this contingency. It is seldom, however, that any such loads as this are hauled even over good roads, simply for the reason that there are usually stretches of poor road to be travelled by the farmer before the good road is reached ; and for many years to come, certainly not within the life of macadam pavements built in the next few years, the average net load will probably not exceed three tons, which would enable a team to climb a short 8% grade without overtax- ing its energies. Since it is generally a few short deep cuts that add so greatly to the cost of a road it would seem to be good en- gineering not to endeavor to reduce the grades where much earth- work is necessary, without careful consideration of the above stated facts. A 2% or 3^ grade forms a desirable maximum if it can be cheaply obtained, since 2% or 3% is the slope of repose of LOCATION f 15 a wagon on a macadam surface, and a horse can in consequence readily trot down such a grade. Having considered the design of cross-section, the cost of exca- vation and surfacing, and the selection of maximum grades or de- sign of profile, let us now briefly consider the subject of the gen- eral location of the road. LOCATION. Upon this subject alone a volume could be writ- ten, but much already exists in print and it is not the purpose of the writer to "re-hash," but rather to call attention to facts which hitherto have not received publication, or, if published, have not become generally known.. Few existing roads are well located. They are built after the style of the Romans, going over hills rather than around them. Any civil engineer would remedy such a glaring defect if given the opportunity ; but there is one point in location that the engin- eer is almost certain to overlook if his experience has been in rail- road work, namely, the character of the soil over which the road runs. Upon one side of a valley the surface may be clay, upon the opposite side gravel ; in the bottom of the valley the soil is usually alluvial, higher up on a bench it is generally far more fit for road purposes. The experienced engineer will therefore not select the final location of a road until he has studied sub-surface conditions as well as topography. In locating a road in a new country the engineer will also bear in mind the fact that one of the greatest items of cost of a gravel or macadam road is the hauling of ma- terial, and will locate his road near the gravel pits or proposed quarry, wherever possible. A few culverts or drains or a few bridges will greatly swell the cost per mile of road, and these like- wise the engineer will avoid if possible ; and where necessary will build them as cheaply as may be, using vitrified pipe or timber for small culverts, instead of masonry arches or iron pipe ; using steel I-beams with wooden floors for bridges under 30 ft. span, instead of the expensive Pony trusses so common in the eastern states. It is not unusual to see highway bridges of 60 ft. or more span over streams that never carry much driftwood or ice (the greatest de- stroyers of piers), where two or more 30-ft. spans resting on pile foundations would suffice, at far less expense. Bridges may fre- quently be avoided entirely by diverting the stream, which is a common and economic expedient in the far west. To sum up this chapter, we find: (1) A uniform cross-section for all parts of the road should not be adopted. The depth of ditches should be made to vary with the character of the soil, very 16 GRAVEL ROADS shallow in sand or on steep grades, and deep in flat, soggy lands, but ordinarily not much more than a foot below the general ground level. The ditches should be wide enough to permit the use of drag and wheel scrapers, and engineers should cease blindly copying plans made long before the invention of the wheel scraper and other labor-saving devices. (2) Freezing of the soil does not destroy a macadam or other road crust, provided there is 1 ft. of dry earth beneath the macadam. (3) The cuts and fills should be balanced, including the earth from the ditches. No sandpapering of slopes and embankments should be specified. (4) Select a profile with a minimum grade of y> ( / f: and a maxi- mum grade of 3%, if it can be obtained at nominal cost ; but up to 5%, or even 8%, if necessary to avoid expensive excavation, for the tractive power of a horse is not a constant quantity and is greater than authorities state. (5) Design a rather flat arch for the road surface, with a total crown or drop of ViH 11 - in 12 ins., and avoid any shoulders or trenches in the cross-section. Thus will it be possible to do the surfacing by horse instead of by man power. (6) Locate the road with due regard to the character of under- lying soil, shifting it where possible to secure better material both for sub-grade and for surfacing. (7) Cut down the cost of all culverts and bridges as low as pos- sible, both by careful location of the road and by economy in the design of culverts and bridges. Chapter III. GRAVEL ROADS. As compared with earth roads, next in point of economy in first cost are gravel roads. As usually constructed they are but little better than earth roads, f or the simple reason that any gravel is considered good enough, and no care is taken to screen out the poor material from the good. More than 30% of sand or loam destroys the effectiveness of gravel, rendering it pervious to water and consequently unserviceable. It may be set down as an almost infallible rule that gravel should be screened, not only to exclude an excess of fines, but to insure an even distribution of fine and coarse material when placed in the road. Where a small amount GRAVEL ROADS . 17 of gravel is required, the ordinary inclined stationary screen, against which the gravel is thrown by shovel, will suffice ; but for extensive work it is cheaper to use a stone-crushing plant with elevator, rotary screen and bins, thus breaking the very large peb- bles, while by storing the gravel in bins the expense of loading with shovels into wagons is avoided, and the team loses very little time in waiting to be loaded. To save cost in handling, the un- screened gravel should be shoveled into small dump cars, hauled on a track up an incline and dumped directly over a chute feeding into the crusher. With eight men loading cars, one horse and driver hauling, one bin man and one engineer, the output should be 150 cu. yds. of all sizes of material in 10 hours, part of which will probably be waste, an excess of fines being usually obtained. With labor at 15 cts. per hour the cost of screening, including coal and engineer, but not rent of plant, will be about 15 cts. per cu. yd. ; since probably one-third of the output will be waste, the cost of useful product will ordinarily be about 20 cts. per cu. yd. in the wagon. The cost of hauling and rolling will be found in the following chapter, being the same as for macadam. The screen should have three sizes of circular openings, % in., 1% and 2% ins. in diameter. Considerable variation may be made from these sizes if experience with the material used gives better results. Engineers, and especially young engineers, are prone to accept all that is published regarding "the proper sizes" and other de- tails of construction without asking the reason why. The writer cannot too strongly urge the necessity of less empiricism and more rationalism on the part of both authors and readers. We are all too apt to believe that the particular way in which we have overcome some difficulty or solved some problem is the best and only right way, when in fact there probably are other and better methods. Even at the risk of being accused of digression the writer wishes to emphasize this contention by an example. For years it has been customary to specify that no stone in a broken stone road shall be over 2Vi> ins. in diameter, because it is claimed that if larger it will work to the surface. There is no doubt that if a mass of loose stone of various sizes is passed over by wheels the larger stones will tilt up when the weight comes upon one end of them and the smaller stones will roll down into the place made vacant by such tilting, and by a repetition of this process the larger stones work to the surface. But it does not follow that in a gravel or broken stone road, rolled with a steam-roller and bound together with the addition of fines, a stone will work to J8 GRAVEL ROADS the top if it is 2 ins. below the surface 1 to begin with. In fact, the mass is so perfectly bound together that it is impossible for tilting to take place. In the city of Rochester, the writer has seen sec- tions of old macadam pavements where stones 5 ins. long are to be found' scattered indiscriminately through the mass, apparently just as they had been placed ten years before. (Jreat as was Macadam, careful as were his observations, and sound as are many of his conclusions, it is evident that in this case his rule that "no stone larger than will enter a man's mouth should go into a road," does not apply; for we use a binder where he did not ; we use a steam-roller where he did not; and these two factors make all the difference in the world in the behavior of the stone forming a macadam pavement. Therefore, let us not be dogmatic in naming limiting sizes of stone to be used in road construction, especially for the lower courses. To return to the screening. The gravel passing through a l 1 /^- in. screen will probably contain most of the softer varieties <>l stone, while the gravel passing through the 2 1 /-in. screen will be of harder consistency ; the smaller-sized stone should therefore form the lower layer of the road, and the larger, tougher sizes the surface. It will require about one load of the screenings (less than Vii-m- in diameter) to every four loads of the larger sizes to fill their voids and bind them together; and the screenings should be spread over the larger gravel, sprinkled, and rolled with a steam-roller until perfectly compact, the method used being much the same as described for macadam in the next chapter. Fre- quently the smaller and rounded gravel cannot be rolled before the addition of the screenings, as it will push before the roller in a wave ; while the top course of stone, if it has any sharp-edged pieces, due to the crushing of larger fragments, may be rolled before the addition of screenings. It has been claimed by some engineers that the rounded form of the fragments of a broken cobble or large pebble make gravel unfit for road purposes, as it tends to rock under a load and so breaks the bond, rendering the road unstable. The writer does not agree with this theory, be- cause even a poor gravel road S!IOW T S no sign of instability ki summer, and, furthermore, a stone thoroughly supported .on all sides by a binder of sand or the like cannot rock. The real cause of the failure of most gravel roads is carelessness in selection of materials, and utter absence of judgment in con- struction, and while a gravel road will probably have to be thicker than a macadam road for equal supporting power, it remains a MACADAM ROADS * 19 fact that good gravel roads have been built even without screen- ing out the excess of fines, and still better ones can be built by following the method of preparation and construction outlined above. The writer must not be misunderstood as arguing for gravel in place of macadam. The latter is always to be preferred as a matter of ultimate economy, where good ledge rock is obtain- able with a moderate haul, for, as will be shown later, the cost of quarrying is usually not to exceed 50 cts. per cu. yd., while the cost of handling and screening gravel is almost as great as the cost of crushing stone, leaving the other items of cost for both macadam and gravel practically identical, if the haul is the same. This would make a difference in cost of about 25/ only, in favor of a gravel road as compared with macadam. Should good stone for macadam surfacing be expensive, and no suitable quarry stone be available even for the lowe'r course, gravel can be profitably used in place of the stone for the body of the road, which may be surfaced with 3 ins. of imported trap or other suitable rock. Telford resorted to this form of construction, but it seems not to have been mentioned in modern text-books. Chapter IV. MACADAM ROADS. The non-professional reader may have wondered why a mass of broken stone, when sprinkled and rolled, finally becomes a solid pavement, impervious to water, acting in all respects like con- crete, although 110 cement mortar has been used. That such a result could be obtained seems not to have occurred to any one before the time of Tresaguet, and even he did not trust to the broken stone alone to sustain loaded wagons, for he used an un- derpinning or "bottoming" of large paving stone. It was Ma,- cadam who, some eighty years ago, by omitting the bottoming, showed conclusively that broken stone possesses the property of knitting together, or becoming cemented under the rolling action of passing Avheels. The writer doubts whether Macadam himself understood the philosophy of this cementing action, and, judged by ^the various explanations offered, it is questionable whether modern engineers have fully comprehended the real cause. WHAT HOLDS MACADAM ROADS TOGETHER ? There are those who say that the roller, by shaking and pound- ing the mass of loose, broken stones placed on a road, finally com- 20 MACADAM ROADS presses the stones together until they are almost, if not quite, as compact as solid rock. This they tell us is the true explanation of the binding under the roller. There are two objections to their theory: (1) The roller does not compress the stone to its original volume; that is, it does not reduce the voids to zero; (2) a road is never bound when the rolling is finished, unless a binder has been added. As a matter of fact, rolling does not reduce the voids of the mass of hard, broken limestone more than one-half, leaving at least 20% voids. This the writer has determined by tests over several miles of road where the output of the crusher was care- fully measured in wagons, and afterward measured rolled in place. As corroborative evidence, the writer refers to the Trans- actions of the American Society of Civil Engineers, for 1899, wherein Mr. W. C. Foster states that in one instance 7.38 ins. of loose trap-rock was rolled to 6 ins., while Mr. Cudworth states that in another instance 3.9 ins. was rolled to 3 ins. It is well known that the voids in loose, machine-broken stone are about 40%, and in order to reduce these voids to zero, 6 ins. would have to be rolled to 3.6 ins. Upon a firm foundation, where no stone can be lost in the sub-grade, and so deceive the experimenter, no one has ever rolled 6 ins. of hard, broken stone to 4 ins., or re- duced the voids to as low as even 6% ; and, common as is the false statement in text-books (and New York State specifications) that this has been or can be done, the statement may be traced to one authority, namely, Macadam, who undoubtedly did see 6 ins. brought down to 4 ins., but it took months to do this under the wearing action of hoofs and wheels, which was not rolling but crushing the stone. Were any further evidence necessary to prove this contention it might be found by observing the volume of screenings or binder necessary to fill the voids in a well-rolled macadam ; and the amount of screenings required is never less than 20% of the volume of the rolled metal, where no screenings are wasted and the voids are completely filled. We therefore con- clude that the interlocking of the fragments of stone does not account for the binding. What then does cause the binding of a macadam road? Evidently the screenings or binding materials are essential. Certain authorities tell us that the screenings have a cementing property akin to the cementing action of quicklime or iron-rust. In other words, that crystallization or setting. takes place. Again the writer cannot agree with the authorities. A true explanation of the phenomenon seems to be found in a study of the sand on a sea beach. BINDING 21 Where the waves break, the sand is firm and makes a very fair road itself, while a little farther back, beyond the reach of the waves, the sand is loose and yielding. The waves have evidently been the means of binding the sand. Each wave as it rushes up on to the beach carries in suspension an amount of fine sand that is precipitated upon the surface of the beach where the wave breaks and is washed down into the voids of the larger particles of sand, thus puddling or filling up all large interstices; but there is one more necessary condition to secure firmness and sustaining power, the sand must be moist. The fine capillary pores hold the water, or rather the water in the pores holds the sand together by virtue of its surface-tension, a well- known physical phenomenon; and it is this surface-tension or viscosity of water that binds the sand together and makes of the sand beach an excellent surface to walk or drive over. Strange as it may seem, dust and water, commonly considered the two greatest enemies of good roads, are, when in their proper places, the two elements that prevent the disintegration of macadam. The writer conceives that the authorities have been wrong in their theories and that macadam is first bound, not by a cementing action, but by the surface-tension of water in the capillary voids of the screenings, and he offers the following facts as evidence of the truth of this theory : (1) A road built without screenings will not bind unless it is left long enough under the action of hoofs and wheels to produce screenings. (2) A road built with screenings will not bind if all the dust has been taken out of the screenings, leaving only the coarser par- ticles, but will bind immediately upon the addition of the dust and water. The writer has tried this experiment under the direction of a "good roads expert" who had ordered all the dust to be screened out, under the mistaken idea that it would be injurious to leave it in ; but, as stated, it was found impossible to bind the road until the dust was spread over its surface and washed into the voids. (3) A road will not bind until sprinkled, even after the screen- ings or binder have been added. (4) Time for iron-rust or other cementing action to take place is not necessary. A newly bound road, one that can be picked to pieces without evidence of any cementation, will uphold a heavy wagon, behaving exactly like an old road. (5) The screenings of a very hard rock like trap bind slowly, 22 MACADAM ROADS sometimes not at all, due to insufficiency of dust necessary to pro- duce capillary voids; but upon addition of a little sand or road- sweepings, bind immediately. Conversely, the screenings of a soft rock, like limestone, rich in dust, bind quickly. (6) Hygroscopic rocks (those that condense moisture upon their surfaces), like limestone, furnish better screenings for binding and do not ravel as quickly in dry weather as siliceous or quartz- like rocks. (7) Long-continued drought causes macadam to ravel and final- ly go all to pieces, while it immediately knits together again after a rain. (8) Macadam in tunnels, where not sprinkled, soon ravels, as do likewise windswept roads that are kept free from sufficient dust and moisture. The writer is not to be understood as advocating the use of dust and water alone to produce binding, rolling being quite as im- portant a factor, if a road is desired that will retain a smooth sur- face. Rolling in the first place so consolidates the stone that there is little chance for play or movement, while the screenings and water added render appreciable movement impossible. Sand and water, or screenings and w r ater, though not a true mortar, serve the same purpose, as we have seen, and the less mor- tar in any masonry structure the more perfect and durable it is. It may be asked w T hat connection the dust-water theory of bind- ing above outlined has to do with economic road construction, for good roads have been built by those to whom such a theory has never occurred, in fact by those to whom the word theory means anything entirely barren of practical results. A true theory is, however, not without economic value. Much time has been spent and much money wasted in vain en- deavor to bind trap-rock without screenings or w r ith screenings containing insufficient dust to make capillary voids. All this would have been saved had the true theory of binding been under- stood. Indeed, it is not uncommon to see it stated that trap-rock screenings are totally unfit for binding, and they are rejected and Tompkins Cove limestone screenings or the like substituted at greater cost. The writer has found that, if upon a road partially bound with trap screenings, a little limestone dust or road sweep- ings is added, the binding is immediately effected. Thus not only is unnecessary rolling in vain endeavor to bind with trap s0 1 helper 1 .50 y% engineer on boiler 1.25 1-3 ton soft coal 75 Repairs to drill and hose 1.00 Total $7.00 With a good drill runner, and in rock that is not very seamy, 70 ft. of hole may be drilled per day at a cost of 10 cts. per ft., but it frequently happens under adverse conditions that this cost is doubled. We shall assume 15 cts. per ft. for drilling. The num- ber of feet of hole drilled per cubic yard of rock varies as the depth of the lift ; or since the depth of hole is about the same as the depth of lift or face, we shall speak hereafter in terms of the depth of hole. It is a common rule to space the holes a distance apart equal to their depth (s=d) ; the writer, however, uses the following original formula for spacing holes in stratified rock, 10 s = yd. 4 Wherein s = distance apart, and d = depth of holes. Comparative results for various depths of holes using the for- 10 mulas, s == d and s - - \/d : 4 Feet of drill hole per cu. yd. of rock. 10 " s = d. 1.0 1.7 0.75 0.40 0.20 While the common rule (s = d) may be a good one in igneous rock, the writer has found that with limestone, sandstone, etc., it does npt produce as satisfactory results as does his own formula. Depth, ft. d. 2' s = > 4 22 4 . 1 1 6 072 8 0.55 12 . 0.36 CRUSHING 27 Q It is evident from the accompanying table that whichever rule is used the cost of quarrying increases rapidly as the depth of hole decreases; whence it is desirable to make the holes not less than 6 ft. deep, and they need not be over 12 ft. to 15 ft. deep to secure economical results. DYNAMITE. The amount of 40% dynamite required per cubic yard of rock excavated varies also with the depth of hole, decreasing as the depth of hole increases. In open cut work the 3 writer uses the original formula, P = , in which P is equal to d the pounds of dynamite required per cubic yard of rock. With 40% dynamite at 15 cts. per It), and drilling at 15 cts. per ft., we find upon summing up that the cost in cents per cubic yard of rock excavation, solid measure, using the writer's rules, is as follows : Where d (in ft.) 123468 10 12 Cost of dynamite per cu. yd. . 26 18 15 13 11 9 8 7 Cost of drilling per cu. yd.... 66 32 22 17 10 9 7 6 Total cost per cu. yd 92 50 37 30 21 18 15 13 Having never seen a rational and detailed explanation showing that the cost of rock excavation varies inversely as the depth of the lift in the quarry, the writer has gone somewhat into detail for the purpose of demonstrating the necessity of opening up a quarry so as to secure considerable depth of face, if it is desired tt) move rock economically. The formulas given are purely empirical, based upon experi- ence, and not upon the wave theory of the effect of explosives, but it is believed that the formulas are rational in construction. Sledging and throwing the stone back from the face will cost about 15 cts. per cu. yd., loose measure, for stratified rock where a 9 x 15-in. crusher is used. To sum up the cost of quarrying, ex- clusive of stripping, pumping and superintendence, we find it to be about as follows: Drilling and dynamite, 21 cts. per cu. yd. solid measure, or about 15 cts. per cu. yd. loose measure, allow- ance being made for waste, to which adding 15 cts. for sledging we have a total of 30 cts. per cu. yd. loose measure for quarrying. CRUSHING. Under this head we shall include the cost of de- livering the stone from the quarry to the crusher and shall assume that a good portable crusher having a 9 x 15-in. opening is used. 28 MACADAM ROADS The output of such a crusher is ordinarily about 60 cu. yds. in ten hours, the stone being measured loose in the bins or wagons and using a rotary screen having three sizes of circular openings, namely: %, 1^4 and 1% ins. The output of the various sizes is about as follows: y 2 -in. stone, 16^ ; 1%-in., 24% 2i/ 2 -in., 60%; total, 100%. Using a stationary screen made of bars % in. apart, the writer has found the screenings to be about 25% of the total. In both cases the jaws of the crusher were set to crush fine, and all material over 2% ins. in diameter was run through the crusher a second time. If the crusher is about 100 ft. distant from the face of the quarry it will take six men with wheelbarrows to supply it with stone, and it need cost no more if the crusher is somewhat further away, provided dumpcarts are used and the stone dumped on the platform. Each man will load and deliver in wheelbarrows about 10 cu. yds. of loose stone per day. It will take two active men to feed this 60 cu. yds. daily into the crusher, at a cost of 5 cts. per cu. yd. for feeding. The engineer will receive about $2.50 a day, equivalent to about 4 cts. per cu. yd. One man will be required to help load the wagons at the bins at 2% cts. per cu. yd., and one man to carry tools to the blacksmith, etc., will be needed at 2% cts. a cu. yd. A blacksmith, sharpening tools at $2.50 a day, a water boy at 50 cts. and a foreman at $3.00 per day add another 10 cts. per cu. yd. to the cost of quarrying and crushing. Summing up, we have the total cost of loose stone per cu. yd. in the wagons, with 60 cu. yds. daily output, as follows : Quarrying and sledging $ .30 Wheeling to platform 15 Feeding crusher 05 Engineer 04 Bin man and tool man 05 Foreman, blacksmith and waterboy .10 2-3 ton of coal at $2.70 03 Wear and tear on plant and interest, $3 day 05 Total $ .77 Delays from breakdowns 08 Stripping, say 15 cts., and quarry rent 10 cts 25 Grand total.. ..$1.10 HAULING AND SPREADING 29 While there is no doubt that this cost may be somewhat reduced, it can be done only where work is upon a sufficiently large scale to warrant the use of a stationary plant of greater size than is ordi- narily used on roadwork. Crushed stone may be brought from the large quarries at 75 cts. per cu. yd., and in certain places where the waste product of a building stone quarry is crushed it may be bought for even less. The writer has seen many absurdly low estimates of the cost of macadam roads made by engineers who have based their prices of stone upon the prices charged by the quarrymen on the Hudson River, where the work is carried on quite differently from what is possible with a portable plant and in a quarry with a shallow face. FIG. 8. MACADAM ROAD WITH SHALLOW DITCHES, AFTER WIN- TER'S USE. (Showing- macadam (12 ins. x 18 ft.) road without ditches; earth shoul- ders, 9 ft. wide; crown of road only 15 ins. above bottom of gutter. Road in perfect condition after severe winter and wet spring.) HAULING. Bins should always be erected to receive the broken stone and so avoid rehandling. A wagon can readily be loaded from a bin in two to five minutes. Ordinarily not more than l 1 /^ cu. yds. can be hauled by a team. The speed of the team going and returning should average 220 ft. per minute. Allowing five minutes to dump the load and five minutes in loading, the fixed cost of hauling is 7 cts. per cu. yd., to which add % ct. per 30 MACADAM ROADS cu. yd. per 100 ft. of lead, or 25 cts. per mile where team and driver are worth 35 cts. per hour. The fixed cost may readily become greater than 7 cts. if there are any delays or breakdowns at the crusher, and it is safe to say, that after figuring the cost by the rule just given, at least W/ ( should be added for such contingencies. It should be borne in mind also that all the teams cannot be loaded at once in the morning, and furthermore that the length of haul is continually changing so that at certain times teams are not working to their full capacity. SPREADING. Two men will dump and spread the output of the crusher, or 30 cu. yds. each, unless the specifications compel dumping on boards and shoveling the material therefrom into the road, under which conditions at least four men will be required. Shoveling from dumping boards is in the writer's opinion a use- less and expensive refinement. It is claimed that dumping a load in one spot on the road results in undue consolidation at that place, but the writer has never seen any evidence of such an effect where the spreader knows his business, and where slat-bottom wagons are used, for the spreader will then not allow the load to fall all in one place, but dump in several small piles, since to do otherwise would make more work for himself. Where the output, of several crushers is daily placed upon the road a Shuart grader may be used to advantage for spreading, as the blade will pick up, push along and level at least 500 cu. yds. per day, requiring a team and driver and one man at a cost of $5, or 1 ct. per cu. yd., and it will cost 1 ct. more per cu. yd. to complete the leveling by hand with a potato hook. The screenings should not be dumped directly upon the broken stone, but placed in piles at convenient intervals along the sides of the road and spread with shovels after the rolling has been almost completed. One man will spread about 10 cu. yds. of screenings per ten-hour day at a cost of 15 cts. per cu. yd. It should have been stated above that screening is necessary to insure the even distribution of the binder throughout the road, and Byrne errs when he advocates the placing of unscreened stone upon a road. Had he observed the way in which the larger stones in a bin roll down and collect at the lower end he would have seen the necessity of separating the sizes by means of a screen to insure even distribution. ROLLING AND SPRINKLING 31 9 ROLLING. Byrne states in his work on "Highway Construc- tion" that in England a roller averages 1,000 sq. yds. of 3-in. fin- ished course of macadam a day, which is equivalent to 85 cu. yds. of macadam rolled ; and he cites one instance where twice this amount of limestone was compacted, the roller passing over it only 35 times. If an excess of binder and water are put on before the coarse stone has been consolidated there is no doubt that macadam may be compacted in the short time above given, but the writer does not believe that a road can be properly built with so little rolling. North states that in one instance where a 6-in. course of trap macadam was applied in making: repairs it took a 15-ton roller 38.2 hrs. per 1,000 sq. yds., or about 44 cu. yds. was rolled per ten-hour day. In another instance it took 58.6 hrs. to pack a 7-in. macadam, laid in two courses, or 33 cu. yds. per ten-hour day. Mr. Cudworth states that a 3-in. finished course of trap was rolled at a rate varying from 38.4 to 65.4 sq. yds. per hour, or an average of about 40 cu. yds. per ten-hour day ; while Mr. Foster states that a 6-in. finished course of trap was rolled at the rate of 31.4 sq. yds. an hour, or 52 cu. yds. per ten-hour day. The writer's experi- ence agrees very closely with that of Mr. North. For trap work, the writer has found that about 36 cu. yds. of macadam can be rolled a day and about 42 cu. yds. where limestone is used. The daily cost of operating a steam roller is about as follows : Engineer $ 3.00 Night watchman 1 .50 4-10 ton anthracite at $5.50 2.20 Oil and waste .30 Interest and depreciation 3.00 Total $10.00 Interest and depreciation, as given, may seem high, but it is a common error among civil engineers not to allow sufficiently for interest and depreciation of plant, because they usually forget that such a plant is not in operation continuously throughout the year. A contractor will seldom average more than 100 working days each season with his roller, therefore, with interest at 6'/v' and depreciation at 6 miles an hour, it is easy to estimate the number of trips a day and the number of sprinklers that will be needed (with varying lengths of haul). Ordinarily one sprink- ler is required for each roller, so that the cost of sprinkling will be 10 cts. per cu. yd., which, added to the pumping, makes a total of 12 cts. per cu. yd. of macadam, but with a long haul and in sandy soil the cost frequently runs as high as 20 cts. per cu. yd. QUANTITY OF MATERIALS AND COST OP WORK.-As stated above, 6 ins. of loose broken stone will compact under the roller to about 4% ins. in thickness, or 1.3 cu. yds. will roll to 1 cu. yd., packed, and the voids will be reduced from about 40% to about 22%, beyond which no amount of rolling with a ten-ton steam-roller will effect further consolidation. This data is based upon careful measurement of loose stone in the wagons and after- wards compacted into a macadam road several miles along 1 . To fill the voids in the compacted stone 0.3 cu. yd. of screening (J/2-in. diameter down to dust) were required per cubic yard of com- pacted macadam. QUANTITY AND COST 33 It might seem at first sight that about 0.22 cu. yd. of screenings would be needed to fill the voids, but an excess must be provided, for the puddling action of the water used in sprinkling, and the crushing action of the roller reduces the volume of the loose screenings not less than 30%. The writer uses this rule, based upon his experience : To ascertain the thickness of the coat of screenings, multiply the thickness of the finished macadam by 0.25 and add 1-3 in. to provide for filling the surface voids and for loss. This rule will be found very accurate for any thickness of macad- am from 2 to 12 ins. To sum up: It requires the following amount of materials to make a 6-in. macadam road, properly rolled and with voids* re- duced to about 4%. 1.3 cu. yds. of ~y<2 to 2 1 /o-in. loose stone, 0.3 " " " M}-i n - and less screenings, 1.6 " " " total loose stone to make 1 cu. yd. macadam, Or, stated differently: 7.8 ins. of loose stone % to 2% ins. in diameter will roll to 6 ins. 1.8 " " " screenings will fill the voids. 9.6 " " " stone and screenings will make 6 ins. of ma- cadam. The reports of the Massachusetts Highway Commission will be found to confirm this data, although their unit of measurement is the ton of 2,000 Ibs., and some assumptions must be made to re- duce to our unit of cubic yards. We shall assume that 1 cu. yd. of trap-rock weighs 11-3 short tons. From the report of 1897 it may be shown that 0.336 short tons of stone and screenings were re- quired per square yard of finished macadam averaging 5% ins. thick, or about 1.7 cu. yd. of loose stone and screenings were re- quired per cubic yard of macadam. The report of 1899 shows that 0.328 tons were required per square yard of macadam, or 1.62 cu. yd. of loose stone and screenings per cu. yd. of macadam. These figures are the average of thousands of square yards of macadam road. It should be stated that in a sandy roadbed or one not perfectly compacted by the roller, % to 1 in. of stone will be pushed into the sub-grade and lost, and if a cushion coat of screenings is specified, that also must be added to the above esti- *Coddrington, In the "Encyclopedia Brittanica," states that voids in well-compacted macadam are 5 per cent, as determined by actual weight. 34 MACADAM ROADS mated quantities. A cushion coat is entirely unnecessary, except to prevent raveling in dry weather, and it is then much cheaper to use sand for this purpose. According to the present New York State specifications all the product of the crusher from % to 1*4 ins. in size, or 25% of the total output, is wasted, since the specifications prohibit its use. This is another of the many absurd extravagances to be found in so-called "standard specifications." while it does not produce a better road, it adds very materially to the cost of roadwork where the contractor can find no market for stone of the prohibited sixe. We are now in position to estimate the cost of a cubic yard of macadam in place, and, assuming the rates above given and that the full product of the crusher is used, find it to be as follows : 1.3 cu. yds. coarse stone at $1.10. $1.4:5 1.3 " " " " hauled 1 mile at 32 cts 42 1.3*' " " " spread at 5 cts 06 1.3" " " " rolled at 20 cts 26 0.3 " " screenings at $1.10 33 0.3 " " hauled 1 mile at 32 cts 10 0.3 " " spread at 15 cts 05 Sprinkling 12 Total, per cu. yd. macadam $2.77 This is for a haul of one mile, and 40 cts. per cu. yd. of macadam must be added for each additional mile haul. As above stated, the amount of screenings will be slightly less than is necessary to bind the road, but it is assumed that a small amount of sand will be allowed to piece out; should this be prohibited by the specifica- tions, screenings will have to be imported at an additional cost. The 1899 report of the Massachusetts Highway Commission shows that the average contract price for broken stone in place was $1.55 per ton, or about $3.40 per cu. yd. of macadam in place during 1898. It may be well at this point to insert, for sake of comparison, the cost per mile of standard road, 15 ft. wide, 6 ins. thick at the center and 5 ins. on the sides, as constructed in the state of Massa- chusetts, where 175 miles of road have been built from 1894 to 1899. The average cost has been about $10,000 per mile of road and the approximate per cent, of each item is as follows ; QUANTITY AND COST 35 Per cent, of total cost. Macadam, at $1.50 ton for local stone, $2.00 ton for trap 55.0 Excavation at 32 cts. cu. yd. for earth and $1.50 for rock 14.0 Engineering 12.0 Culverts at $100 each and bridge at $1,000 each. . 6.0 Gravel for shoulders at 60 cts. cu. yd 5.0 Shaping or surfacing at 2 cts. sq. yd 2.0 Side drains at 29 cts. lin. ft 2.0 Stone bounds and miscl 2.0 Guard-rail at 13 cts. lin. ft 1.5 Telford bottoming, at 31 cts. sq. yd 0.5 Total 100.0 The most striking fact brought out by this tabulation is that the macadam itself has cost little more than half the total, while the engineering alone has been 12% and the earthwork 14%, with sundry items forming the balance. What shall be said of such road construction ? However excellent it may be in point of artis- tic finish and utility, a road built at such a cost indicates an ex- travagance that would not be tolerated by a private corporation just entering upon the construction of good roads. The item of earthwork is excessive, both as to quantity of earth moved and as to cost per cubic yard ; but we shall not here repeat the criticism made in Chapter II. of the design of road cross-section that has caused so great an increase in this item. The item of shaping, as we have shown, is likewise greatly in ex- cess, of what it would be were the use of graders possible. Culverts and bridges should ordinarily cost not half as much per mile of road as above given; while guard-rails of the expensive design specified are uncalled for on a new road, and as a matter of fact guard-rails are generally quite unnecessary, for while a driver might occasionally go over an embankment, were the horse pos- sessed of no intelligence, the latter condition seldom exists. As to the cost of engineering and inspection, we find this item to be about three times as great as it should be. The writer is an engineer and would not knowingly advise the rejection of an en- gineer's services on the ground of high cost. The fact is that no profession is so poorly paid, considering the brains and ability 36 MACADAM ROADS represented ; but the man of wisdom will see that if the roads can be built more cheaply, more roads will be built, and the engineer's services will be more in demand. The mere survey, estimate and plans should not, and usually do not, cost over $50 per mile ; while the staking out of the work need not cost over $30 per mile. It is therefore evident that inspection forms the greatest portion of the engineering item, and this is in consequence of the slow progress made by a contractor with a small and inadequate plant. The contractor should be required to build one mile of road a month, once he has installed his crushing plant and got his quarry stripped and ready to operate, a reasonable allowance always be- ing made in case of bad weather. With an engineer at $5 a day and a helper at $3 a day the cost would then be about $250 per mile of road for inspection, or a total cost of $330 for survey and inspection ; but since the engineer and helper cannot work contin- FIG. 9. ANOTHER EXAMPLE OF SHALLOW DITCHES. (Showing macadam (6 ins. x 16 ft.) road; no ditches; earth shoulders 5 ft. wide; crown 9 ins. above gutter. Macadam perfect after severe winter and wet spring-.) ' " *PL - ually, and yet should not be laid off or dismissed between time, it is safe to say that engineering and inspection will cost $500 per mile. MACADAM ROADS 37 By way of comparison we shall here give the. cost of a mile of road built according to Fig. 7, with all unnecessary items ex- cluded. 3,000 cu. yds. earth excavation at $0.20 $ 600 1,500 cu. yds. macadam in place at $3.50 5,250 Culverts 300 Miscellaneous 350 Engineering and inspection 500 Total $7,000 This price is just about two-thirds what it is costing the State of Massachusetts, but the writer is not merely theorizing, for he has constructed roads over bad country for less money than he has named above by rigorously excluding all unnecessary items, such as stone-bounds, cast-iron pipe, expensive little bridges and cul- verts, and by the use of rational specifications and cross-sections. In regard to the cost of engineering he can also speak authorita- tively, for he has laid out more than thirty miles of road himself where the engineering and inspection have together cost less than 6% of the total cost of the work. In closing this discussion of costs the writer should in justice state that the Massachusetts roads have been built in the worst sections of the state and in small contracts of about two-thirds of a mile in length at a time, both of which tend to increase the cost of engineering and other items. Economy in construction depends very largely upon the length of road to be built, for it costs as much to move and set up a plant to do half a mile of road build- ing as for two miles. The fixed cost of opening a quarry and get- ting roads in shape over which to haul is likewise almost as great for a small piece of work as for a larger piece. To build roads economically it is therefore necessary to construct several miles at a stretch, and it is also necessary to build several thousand dollars worth of roads each year in each state, for thus only can contract- ors afford to provide themselves with adequate plant and tools with which to do the work cheaply. A road-construction plant is more expensive than is generally known, the cost of a complete porta'ble plant being about $7,000, distributed as follows : 38 TELFORD ROADS 1 crusher, 9 x 15 ins., with rotary screen $1,000 Portable bins 200 1 15-HP. engine 200 1 20-HP; boiler 600 12 wheel scrapers 500 12 drag scrapers, shovels and picks 100 2 Shuart graders 100 2 steam drills 500 1 5-HP. boiler for drills 400 Water and steam-pipes, quarry tools, etc .* . . 300 2 sprinkling wagons 500 1 10-ton steam-roller. . 2,500 Total $6,900 Due to the fact that few towns can afford such a plant, they en- deavor to get along with the crusher alone, with the result that they neglect proper grading, do no rolling of sub-grade at all and waste half their stone by its becoming mixed with earth ; use field stone, often rotten to begin with, because they cannot afford steam-drills and quarrying plant ; have a portable crusher and no bins, thereby rehandling their stone; and finally, which is the worst of all, they possess no steam-roller, without which it is im- possible to secure economic road construction. So much twaddle has been published about the desirability of each township or village owning a crusher that there are scores of crushers to each steam-roller in the State of New York, whereas there should be one roller to every crusher. We pass now to a brief consideration of a type of road formerly very popular and still adapted to special conditions. Chapter V. TELFORD ROADS. A Telford pavement consists of an underpinning or bottoming of large stone, usually not less than 6 ins. nor more than 12 ins. deep, set on edge like a rough block pavement and supporting a layer of macadam or broken stone. Telford seems to be especially adapted to wet soils, not easily drained, where small broken stone would be pushed down into the soft soil, becoming mixed therewith. Wherever a cut is made, TELI^ORD especially where the grade is fiat and in a clay soil, or one of quicksand nature, it is always advisable to use the telford construc- tion instead of macadam, for the large underpinning stone form an excellent drain and one that does not easily clog up, like pipes or side ditches, in freezing weather. In the writer's estimation telford is far preferable to macadam; also, in city or village streets where side ditches cannot be built and where in consequence there is usually an amount of ground water in the soil directly under the pavement sufficient in most cases to keep the sub-grade in a more or less yielding condition. In Rochester, N. Y., a telford pavement, consisting of an under- pinning 10 ins. deep, of local limestone or sandstone, surfaced with 4 ins. of broken trap-rock, costs about $1.25 per sq. yd., and is a pavement that is becoming deservedly popular. There are, never- theless, many side streets where a 6-in. course of macadam alone would be about half as expensive and equally as satisfactory, where the teaming is not heavy. It has been stated in a widely read pamphlet on roads and pavements that neither macadam nor telford is an economic pavement for the streets of a village or city. The reasons cited in support of this statement are purely theoret- ical and based upon insufficient data to warrant so sweeping a con- clusion, for not only in the city of Rochester, but in scores of other cities, are to be found most excellent examples of street pavements made of broken stone. There are, it is true, many macadam pave- ments that have gone to pieces under heavy traffic, but the same may be said of brick and of asphalt where improperly constructed or where poor materials have been used. The telford construction may be used to advantage where a crusher is not available, broken stone being imported for the surface coat and local stone used for the underpinning. As to the cost of a telford pavement, the under- pinning will usually cost about 70 cts. per cu. yd. corded up in the quarry and will shrink in laying about 20%, bringing the cost for the stone alone to 85 cts. per cu. yd., to which must be added about 25 cts. per mile for hauling, and 25 cts. per cu. yd. for laying, mak- ing a total of $1.35 where the haul is one mile. To this add 10% for foreman and 15%) for contractor's profits, giving a total of $1.70 per cu. yd. in place. Where the haul is two miles the cost will be about $2 per cu. yd., or about 33 cts. per sq. yd., for a 6-in. course of such underpinning. The average price paid in Massachusetts in 1898 for some 10,000 sq. yds., 6 ins. thick, was 31 cts. per sq. yd. for this underpinning or bottoming. It is to be observed that a 6-in. course of under- 40 REPAIRS AND MAINTENANCE pinning can be put in just as cheaply as a 4-in. course of macadam, with a haul of two miles. A telford road is therefore not as ex- pensive as might seem at first sight. Chapter VI. REPAIRS AND MAINTENANCE. CONTINUOUS VS. INTERMITTENT SYSTEM. While Eng- lish and French authorities on road construction advocate the con- tinuous system of repairing as being more economical than the in- termittent system, the fact that conditions under which we work may be different from those from which their experience was gained, due measurably to the use of rock-crushers and steam- rollers, certainly leaves the question an open one. The intermittent system prevailing abroad seems to consist in loosening the old crust of a road, which has been allowed to be- come badly rutted, with hand picks, and then spreading over a thin layer of new metal which is rolled until bound. It is safe to assume, in the absence of cost data, that this method is an unduly expensive one to employ. The following data, taken from the author's time-books, give the actual cost of resurfacing a macadam road over a mile in length. While the road was worn unevenly, but little new metal (hard limestone) was needed. A 12-ton Buffalo Pitts roller was used, which was provided with steel picks on the rear wheels, and 80 hours of rolling were re- quired to break up 19,400 sq. yds. (240 sq. yds. per hour) of the crust, which w r as exceedingly hard. The picks simply opened up cracks in the crust to a depth of about 4 ins. and it was necessary to follow the roller with a gang of laborers with hand-picks to complete the loosening process. The labor of loosening and spread- ing anew the metal was slightly in excess of 10 sq. yds. per man- hour, 40% of which labor was used in respreading with shovels and potato hooks. After respreading, a short section was so thoroughly drenched by a sprinkling cart, that when the roller came upon the metal, the screenings, which had settled to the bottom during the spread- ing process, were floated up into the interstices. The roller and sprinkler were in use but 63 hours, or 300 sq. yds. per hour. This rapid rate was due to the large amount of water used (the water haul being short), the unyielding telford foundation beneath, the SANDSTONE MACADAM 41 Jj abundance of screenings and fine dust, and the use of a heavy roller operated at a high speed. Summing up, the costs are seen to have been as follows : Cts. per sq. yd. Picking with roller at $1.00 per hour 0.4 Picking by hand labor at 20 cts. per hour 1.2 Respreading with roller at $1.00 per hour 0.8 Rolling with roller at $1.00 per hour 0.33 Sprinkling with cart at 40 cts. per hour 0.13 Foreman, 143 hours at 30 cts., for 19,400 sq. yds. . . . 0.44 Total 3.30 At this rate a country macadam road 16 ft. wide can be resur- faced for little more than $300 per mile, and, if built of hard lime- stone, will last on the average for five years. It is difficult to see how any system of continuous repair, with the inefficient methods necessarily employed, can be as economical as work done in the manner above described. It should not be understood, however, that an entire neglect of roads between re- pair periods is favored, for, at times of heavy rains and snows, ditches and culverts require attention, and there should always be someone whose duty it is to look after such matters. SANDSTONE MACADAM. Sandstone has usually been pro- nounced by authorities on highway construction to be unfit for road surfacing. It has, however, been successfully employed for that purpose in Albion, N. Y., a city of 5,000 inhabitants, and the experience there gained may be a valuable guide to others. The average width of the residence streets between curbs is about 28 ft. ; of this, 18 ft. is paved with 6 or 8 ins. of sandstone macadam, leaving a 5-ft. earth shoulder and gutter on each side. A very sharp crown is used, about 1-in. per ft., or a center rise of 14 ins. above the gutter in a street 28 ft. wide. After a heavy rainfall the earth gutters become covered with a thin layer of fine materials evidently washed off the macadam. Hitherto the author has advocated a rather flat crown, about % i n - per ft., in order not to concentrate the traffic at the center of the road. But in a village street, especially if the macadam is of stone that wears rapidly, it is evidently far preferable to resurface the street more frequently as a consequence of giving it a sharp crown, than to have a dusty 42 REPAIRS AND MAINTENANCE or muddy street, which, costing considerable to keep clean by hoe- ing, is generally left until it becomes positively unbearable. It is true that the material washed by rains into gutters must eventual- ly be removed, but this is much more cheaply done at long inter- vals when greater quantities are to be handled. Moreover, since the gutters in question are of earth, a Shuart or other leveling scraper can be used to gather up the detritus and deposit it in piles for removal; whereas such a scraper loosens the surface of a macadam road -and should not be used for cleaning macadam. On country roads made of hard rock and swept by winds, a flat crown is desirable, but where the conditions are as in Albion, a very sharp crown is evidently to be preferred. Medina sandstone by the rattler test has a coefficient of wear about the same as Rockland Lake trap, whence it might be infer- red that they would exhibit practically equal wearing qualities in a road. There is, however, a much greater uniformity of hardness in trap, which insures a more even surface wear under traffic than is to be found in sandstone, which, like brick, appears to give way in spots that rapidly grow larger. The rattler test fails to show one great source of wear in a road, namely : the crushing of loose fragments that, having been picked out by horses' calks, are ground to powder under passing wheels. This source of wear is greatest where the binding quality of the rock is least, thus plac- ing a hard rock like trap more nearly on a par with a soft lime- stone than the rattler test would indicate. Whatever laboratory tests may show, it is certain that the Medina sandstone on the main street of Albion, under a heavy traffic, has lasted 5 years without resurfacing, and is only (at the time of observation) just reaching an uneven condition that renders resurfacing necessary. Several of the residence streets are still in excellent condition after 5 years' wear. In this connection it is well to remember that because trap-rock wears twice as long as limestone, for example, it by no means fol- lows that it will pay to use trap even if it costs double what lime- stone does, as the following table clearly shows: Annual charge per sq. yd. in cts. Annual Interest New ma- Material, wear. at 5% terial. Total. 6-in. limestone macadam 0.5 in. 1.5 2.5 4.0 6-in., trap macadam 0.1 in. 3.0 1.0 4.0 2-in. limestone macadam 0.3 in. 0.5 1.5 2.0 2-in. trap macadam 0.1 in. 1.0 1.0 2.0 REPAIRS AND MAINTENANCE 43 9 Note. Limestone macadam assumed at $1.80 per cu. yd., trap macadam at $3.60 per cu. yd. From the table it is evident that in a 6-in. road, limestone may wear five times faster than trap, prices being as given, and yet be no more expensive annually; while if a 2-in. surface coat is used, limestone must wear more than three times faster than trap to be less economical, at the cost and rate per cent, assumed. In the method of construction employed in Albion no attempt is made to grade to any established line. Mud-holes are cleaned out and stone thrown in to fill up to the general sub-grade level (as a general proposition, hard earth crusts, the product of years of rolling and compacting by wagon wheels, should not be broken up merely for the sake of securing a uniform grade). The sand- stone is then rolled with a 15-ton roller, and( being hand-broken stone containing no screenings, is more or less broken into smaller fragments by the roller. When rolled solid enough to support a wide-tired wagon without displacement of the rolled stone, wag- ons loaded with sand and loam are driven on. These are spread with shovels in a thin layer and washed down into the voids with a sprinkling cart or hose. After the surface is dry enough so that it will not stick to the wheels, it is again rolled. Spreading of sand, puddling and rolling are repeated once more before the sur- face is satisfactory. In resurfacing, a heavy harrow, similar to those used on farms, is employed for completing the loosening of the macadam after it has been cracked open by the spikes in the rear roller wheels. This is constructed from 4-ins. x 5-ins. oak pieces, 5 ft. 6 ins. in length, provided with pointed iron teeth, 1-in. in diameter, and spaced about 10 in. apart ; these teeth project about 6 ins. below the wood- en frame. This harrowing process not only completes the break- ing up of the crust as well as it could be done by men using picks, but in addition the teeth spread the loosened stone, filling up the low places. The total cost of resurfacing is given in the following table, and is considerably under the figures given at the beginning of the chapter : Cts. per sq. yd. Roller and engineer at $1 per hour, picking 0.5 Roller and engineer at $1 per hour, re-rolling 0.5 Sprinkling with cart, 40 cts. per hour 0.2 Harrowing with team and driver, 30 cts. per hour. . 0.3 Total.. 1.5 44 ROAD SPECIFICATIONS At this rate a macadam road 16 ft. wide and a mile long can be resurfaced for less than $140, or about one-half the cost cited earlier, less than $30 per mile per annum. In addition to this labor cost some 75 cu. yds. of stone are required which cost a trifle over $60, delivered and spread, making- a total of $90 per mile per an- num for labor and material for resurfacing a Medina sandstone road. This particular stretch of road was a main-traveled street and the loss by wear was even less than the amount of new metal used in repairing, about 0.2 in. per annum over the whole sur- face. Chapter VII. SUGGESTED IMPROVEMENTS IN EXISTING ROAD SPECI- FICATIONS. The author would call particular attention to certain clauses in the "standard specifications" for macadam work as exemplified in present practice in the state of New York and Massachusetts, which, in his opinion, are open to criticism. The clauses to be taken up are quoted from the specifications for recent New York State roadwork; equally indefensible clauses are to be found in those of Massachusetts. The first paragraph is as follows : KIND AND SIZES OP BROKEN STONE. 'The broken stone shall be of two courses. The bottom course will be 4 ins. thick and may consist of trap-rock, granite or any of the harder grade of limestone not inferior to the specimens exhib- ited in the offices of the State Engineer and Surveyor and of the Engineer of the Western Division, which stone shall be broken in sizes varying from a minimum of 1% ins. to a maximum of 3 ins. in their longest dimensions. "The top course shall be 2 ins. thick after rolling, and shall con- sist of trap-rock broken in sizes varying from a minimum of 1 in. to a maximum of 2 ins. in their longest dimensions. "Limestone screenings not-exceeding % in. in size and free from all dirt shall be added to fill all interstices that cannot be filled by the rolling or compacting of the other stone. Such screenings must be free from earth, sand, loam or vegetable matter and shall contain all the dust of fracture." ROAD SPECIFICATIONS 45 t It will be noticed that no provision is made for the use of the % to IV-i-in. sized product of local stone in the lower course. This size of stone forms about 25 % of the crusher output, and its rejec- tion, therefore, means an added cost of 33 c / ( per cu. yd. of product that can be used. That this waste is entirely unjustifiable can be proven by the existence of good roads where every particle of stone coming- from the crusher has been used. Secondly, it will be seen in the specification that the binder for both courses must be limestone. This, no doubt, arises from the assumption that quartz-like screenings will not bind. Sandstone has, however, been used in many roads as a binder, and where slow in binding, the addition of a little dust, preferably limestone, quickly accomplishes the desired end. Trap-rock screenings are also excluded from use by the specification. Since trap is not a local product, but is shipped from the Palisades by boat or rail, it might seem at first sight that the State incurs no loss in specify- ing limestone screenings for a binder, as Tompkins Cove screen- ings certainly bind, more quickly and cost but little more than trap. But on inspection it will be seen that if trap screenings are uniformly rejected, the price of the coarse trap must ultimately be raised for the same reason as above given regarding the waste of middle-sized local stone. Trap screenings, on account of the small amount of dust they contain, bind very slowly, and not until enough dust has been produced by the grinding of the coarse stone under the roller. This grinding, however, wears the coarse stone round, and is therefore objectionable. To cause trap screenings to bind quickly, simply add a little dust, preferably road-sweep- ings or limestone. All difficulty from using trap screenings will then vanish, and their use will cheapen road construction. As to the size of trap specified, the author has been unable to find any quarry man on the Hudson who uses screens that produce stone "not less than 1 in. and not more than 2 ins. in diameter." This specification, however, is practically disregarded, most con- tractors using the "2%-in. stone," ranging from 2 l / to 3 l /4 in. in diameter, and its use has proved very satisfactory. DEPTH OF PAVEMENT. "The pavement, when completed, shall be at least 6 ins. in depth, irrespective of the surfacing material, as is required by the speci- fications, and of such crown and form of gutter as are shown on plans ; and in any case the thickness of the pavement is to be de- termined on a line at right angles to the grade and crown. 46 ROAD SPECIFICATIONS "NO ALLOWANCE WILL BE MADE FOR ANY MATERIAL DRIVEN INTO THE SUBGRADE BY ROLLING OR MISTAKE MADE BY CONTRACTOR IN EXCAVATING OR FILLING. The use of a proper roller, rammers or other suitable implement is to be substituted for that of the steam-roller when the engineer so directs. ' ' Here is seen, what is not uncommon, an effort to throw all the uncertainty of cost upon the contractor, who must guess whether there will be 1 in. or 4 ins. of stone driven into the subgrade, and if his guess is not correct, he must lose or gain as the case may be. Such specifications are not only morally but economically wrong, for the inexperienced contractor is almost certain to lose a just compensation for work for which the State is able and willing to pay, while the experienced contractor will err, if at all, on the side of safety, and so receive pay for material that he will not have to furnish. This specification is, moreover, based upon a false reason, for it implies that all stone driven into sub-grade is lost and valueless, which is not the case. When stone is driven down, the soil is forced up into the voids, and there results a macadam with a binder of sand or loam, which, if not as good as macadam bound with screenings, is nevertheless capable of aiding in the distribu- tion of a concentrated wheel load, and therefore performs the most important function of a pavement. The only fair way to pay for macadam in thin roads is either by the ton or by the cubic yard in the wagons. No effort should be made to finish a road surface exactly to a predetermined grade, as is the present practice in Massachusetts and New York. FINAL SURFACING. "After the wave has been produced over the whole section of the road, screenings shall be again spread on where required to leave them %-iii. deep for a wearing surface. ' ' To cover a 16-ft. road to a depth of % in. for a mile requires about 100 cu. yds. of screenings, which, when imported by rail and hauled, seldom cost less than $2 a cu. yd., and frequently $3. Jhus, to provide a perfectly useless "wearing surface" we find the State expending some $200 to $300 a mile. Any screenings ove"r and above those needed to fill voids serve no useful function upon a ROAD SPECIFICATIONS 47 road except to prevent .horses' hoofs from loosening the stones, and to keep the macadam from raveling, for which purpose sand or loam serves equally well. MATERIAL FOR EMBANKMENTS. "Embankments shall be formed of clear earth or other mater- ials satisfactory to the engineer, and shall be free from vegetable matter or refuse of any kind. ' ' In order to carry out this provision 2,000 cu. yds. of soil con- taining sod are removed and wasted per mile of road, which means a waste of about $500 per mile. The roots of grass extend the full depth of a plow furrow, and their removal can only be accom- plished by taking with them the earth to a depth of 6 or 8 ins. While sod may not be desirable in a very shallow embankment covered by macadam, it may to advantage be used to form the shoulders at the sides. Green grass and weeds may be economical- ly and efficiently removed by using a mattock and taking off only about I 1 /-; ins. of soil instead of 6 or 8 ins., cost '% ct. per s<|. yd. of surface, with labor at 15 cts. per hour. THICKNESS AND WIDTH OF PAVEMENT. The standard thickness of the macadam on New York State roads is 6 ins., and the width 16 ft., with occasionally a 12-ft. road. Upon very soft sand or clay subsoil it is necessary to have 6 ins. or more of pavement, but upon ordinary, loamy or gravelly soils 4 ins. will suffice. The main function of a pavement is not to shed water, as commonly stated, but to distribute a concentrated wheel load over a sufficiently large area of subsoil that will not yield un- der pressure. Incidentally, the drier the soil (sand excepted) the more resistance it gives to pressure, whence it is desirable to pro- vide a pavement that will keep the subsoil dry, a secondary though important consideration. A thick pavement is necessary upon sand, and also upon clay when the latter is not well drained; but this thickness should not always be obtained by the use of expensive broken stone, but by gravel where available at less cost. Thus, 3 to 6 ins. of gravel should ordinarily form the lower course upon which 3 or 4 ins. of macadam should be laid for a wearing coat, instead of the invari- able 6 ins. of macadam prescribed in New York and Massa- chusetts. 48 SUMMARY AND CONCLUSIONS For several reasons the author inclines to the opinion that a nar- row road (8-ft.) would compare very favorably with a wide (16- ft.) road. Traffic follows the center of a 16-ft. road in any case, and the wear is practically as concentrated as on a narrower road, while the loss due to weathering is only one-half as great upon an 8-ft. road. One objection to an 8-ft. road is that two teams can- not pass on it, but this difficulty is successfully disposed of, where the road is built for the hauling of farm produce and not for a speedway, by providing turnouts, 40 ft. long, every 320 ft. These turnouts, having a width of 16 ft., are equivalent to adding 1 ft. of width to the road, and a 9 ft. macadam pavement 4 ins. thick would cost only $1,800 a mile, with stone at $3.00 per cu. yd. By following the foregoing suggestions it is perfectly feasible to build good, serviceable macadam roads in New York State for about $3,500 a mile. They will not be speedways upon which two teams can pass at a full trot, but they will be hard, unyielding surfaces over which great loads may be drawn in all kinds of weather. Chapter VIII. SUMMARY AND CONCLUSIONS. At the close of a discussion it is always well to sum up the cer- dinal points which have been brought out in order to impress upon the mind conclusions worthy of remembrance, if any there are. In the second chapter the fact was brought out that economic road construction has not been the first consideration of engineers em- ployed by certain states, however excellent otherwise may have been the structures designed by them. It was there shown that to perform earthwork economically the engineer must understand the efficiency of modern tools that can be used, and so design the cross-section of a road as to make such use possible. We have indicated the possibility and the desirability of reduc- ing the width of cuts by the use of sub-drain, and the necessity of keeping the mouths of such drains open after a snow. We have shown that deep cuts may be avoided by a considera- tion of the fact that the power of a team is not constant, but may be as great as 1,000 Ibs. tractive pull for a short time, sufficient to mount a grade of 7% with a net load of four tons. We have pointed out the fact that the heaving action of frost does not ma- terially affect a macadam road that has a foot of dry soil beneath SUMMARY it, and that the excessive depth of ditches so commonly seen is un- called for. In Chapter III. attention was called to the necessity of screen- ing all gravel that is to be used in road construction, in order to secure an even distribution of coarse and fine, as well as to ex- clude an excess of fines. It was shown that large stone do not work to the surface of a gravel or a macadam road where a binder has been used, and that those who assert that such an action takes place do so entirely upon the authority of Macadam, who used no binder at all. In Chapter IV. the endeavor was made to establish a new theory as to the cause of the binding that takes place when a broken stone is rolled and sprinkled ; and, whether the theory that water in the capillary voids of the screenings is the true binding agent shall be found to be a complete and perfect theory or not, we have at any rate shown the necessity of having fine dust in the screen- ings or binder. We have indicated in detail the places where the expenses of constructing a macadam road may be legitimately reduced, begin- ning with less earthwork, the use of inexpensive binder, the cut- ting out of many expensive culverts, bridges and miscellanies, and finally ending with a decided reduction in the cost of engineering. We have suggested the construction of all roads by contract upon a scale sufficiently large to warrant outlay for a good plant ; since a good road at a reasonable price has not been built by a state, town or municipality under any other than the contract sys- tem. The reason for this is that each road district or town cannot usually afford the plant necessary, and even though it could, the efficiency of men working for the public is well known to be far below the efficiency of the same men working for an individual. In Chapter V. it is pointed out that the telf ord construction is not as expensive as is commonly supposed, and that it is particularly adapted for cuts and in places where side ditches cannot be made, or where if made, are liable to become clogged with snow. An article upon road construction would be incomplete without some reference to the laws under which roads are built. A good State law wherein the State at large bears one-third the burden, the town or city or county directly benefited bearing also one- third, the taxpayers along the line of the road bearing the re- mainder, is probably the most satisfactory law under which to construct roads. Thus, any farmer, .who owns a team and desires work, can pay his share of the taxes several times over from the 50 j/c V 8T.1MMARY 4N,D ( CONCLUSIONS money he earns, securing all the benefits of the old system of working out taxes, with none of the ills. The great popularity of such a law is to be found by observing its operation in New York, Massachusetts, New Jersey, Washington and other states. In New York State the new law, but two years old, has brought out petitions for over 400 miles of macadam road, which it will take four years to complete, even with an expenditure of $1,000.000 annually, while the object lesson given by the construction of these roads will result in petitions for hundreds of miles more. A good macadam road makes it possible to reach market in all kinds of weather and at a speed before unthought of. It brings the city man into the country, and by creating a demand for prop- erty for residence purposes, greatly enhances the value of the farms. A good road makes possible the hauling of loads weighing three to five tons, and it reduces the wear and tear of wagons, harnesses, horses and men. Road construction puts money into circulation where the road is being built, fully three-fourths of the money coming from out- side sources and remaining in the road district. No farmer with a grain of intelligence will oppose road con- struction under a good law in the face of all these facts ; and, to the credit of the shrewdness of the American farmer, be it said that in no community where roads have been built as described, has the verdict been anything but unanimous in favor of more good roads. ~^ s=:::=:: ^ DAV 0V (G936S) c/u UNIVERSITY OF CALIFORNIA LIBRAR