Department For International )FID ~~~Development TITLE: by: TRL Limited Crowthorne Berkshire RG45 6AU United Kingdom The Location and Engineering Properties of Volcanic Cinder Gravels in Ethiopia D Newill and Kassaye Aklilu PAl 100180 NEWILL, D, AND ASSAYE AKLILU (1980). The location and engineering properties of volcanic cinder gravels in Ethiopia. In: Seventh Regional Conference for Africa on Soil Mechanics and Foundation Engineering, Accra, June 1980. Seventh Regional Conference for A frica on Soil Mechanics and Foundation Engineering /Accra IJune 1980 The location and engineering properties of volcanic cinder gravels in Ethiopia D.NEWILL Transport & Road Research Laboratory, Crowthorne, Berks., UK KASSAYE AKLILU Ethiopian Road Authority, Addis Ababa, Ethiopia Volcanic cinder gravels occur extensively in Ethiopia, but in the past they have been used for road construction only to a' limited extent, even though their use would sub- stantially reduce road construction costs in many instances. As part of a joint research project undertaken by the Ethiopian Road Authority and the United Kingdom Transport and Road Research Laboratory, research has been carried out to provide infor- mation on the occurrence and properties of the volcanic cinders with the object of encouraging their wider use in future road construction. This paper describes the procedures used to map the occurrence of cinder gravels in Ethiopia, reports the results of laboratory tests made on a range of the materials, and gives details of an investigation of an existing cinder gravel road. With the aid of aerial photographs a map was prepared showing the distribution of cinder gravels throughout Ethiopia. The laboratory investigation showed that cinder gravels, which typically have weak particles and are deficient in fine material, are improved by cbmpaction in that some breakdown of the larger particles occurs, producing a better grading and higher 'strength. It was also found that the addition of fines improves the strength and density of the compacted cinder gravels. The road examination confirmed that the action of traffic produces the same breakdown effect in gravel roads as was observed in the laboratory compaction tests. RESUME L'Ethiopie contient des quantit~s consid6rables de graviers de sconies volcaniques, mais, m~eme si leur utilisation peut, dans de nombreux cas, produire une r~duction substantielle des co~its de construction routi~hre, ils n'ont 6t6 utilisgs que de fa~on limitge dans la construction des routes. Dans le cadre dtun projet de recherche entrepris conjointement par les Autorit~s des Routes d'gthiopie et le Transport and Road Research Laboratory du Royaume-Uni, des recherches ont kg effectuges pour fournir des informations sur la prgsence et les proprie'tgs des sconies volcaniques, en v-te d'encourager dans 1'avenir une utilisation plus large de ces derni~res en con- struction routi~re. Ce rapport decrit les proc6d~s utilis~s pour produire des cartes montrant llem placement des scories volcaniques en Ethiopie, pr~sente les rAsultats d'essais en laboratoire effectu6s sur un certain nombre de ces mat~.iaux et donne des d~tails sur des recherches faites sur une route d~ja construite contenant du gravier de sconies volcaniques.Des photos a~riennes ont &t employees pour pr~parer une carte montrant la distribu- tion des graviers de sconies dans toute l'Ethiopie. Les recherches en laboratoire ont prouve que les graviers de sconies, qui, typiquement, contiennent des grains de resistance faible et une quantitg insuffisante de mat~riaux fins, sont amgliorgs par compactage, du fait que les grains les plus larges sont partiellement brise's en morceaux plus petits, produisant ainsi une granularity amgliorge et une resistance plus e'levge. Les essais ont aussi montrg que lladdition de'~lements fins angliore la resistance et la densit6 des graviers de sconies compactgs. Les observations faites sur la route ont confirm6 que l'action du trafic produit dans les routes en gravier le m~me effet de fragmentation que celui observ6 dans les essais de compactage en laboratoire. 21 1 INTRODUCTION In 1975 a joint research project was com- menced in Ethiopia by the Ethiopian Road Authority and Transport and Road Research Laboratory (UK). One of the objectives of the research programme was to establish guidelines to enable highway engineers to make fuller use of locally-available materials for road construction. This was considered to be particularly relevant in Ethiopia as, in common with many other developing countries, an extensive pro- gramme of rural road construction is in progress which is expected to continue for a number of years. For rural roads, which generally carry low volumes of traffic, earth or gravel roads are usually an adequate form of construction. In the case of gravel roads an important factor in helping to reduce costs is to make the best use of locally- available materials. Uusally these materials are natural gravels which will often vary in quality from one deposit to another. The engineer requires guidance in the selection of natural gravels for road construiction, and specifications for their use appropriate to the types of materials available, and the service con- ditions anticipated in the road. This paper describes an investigation of one type of material, volcanic cinder gravel, which, although widespread in Ethiopia has been used for road construction only to a limited extent. One reason for the limited use of volcanic cinder gravels up to the present is that they are generally deficient of fine material and do not conform with the grading specifications for conventional crushed rock bases. Another reason is that they have a reputa- tion for being difficult to compact. The study, aimed at a full assessment of cinder gravels for road construction in Ethiopia comprised the following stages: i) A field survey to locate and identify cinder gravels and to obtain samples for laboratory testing. The survey als'o included the examination of aerial photo- graphs and photo-mosaics and the~prepara- tion of a map showing the distribution of cinder deposits throughout Ethiopia. ii) A laboratory investigation to determine their physical and engineering properties. iii) An examination of existing cinder gravel roads to compare the performance of the materials under trafficking with the results of the laboratory investiga- tions. iv) Compaction trials to determine the most suitable plant for practical use. v) A full-s cale experiment to examine the behaviour of different cinder gravels under controlled conditions in relation to traf- fic and climate. This paper describes the first three stages of the investigation. 2 DEFINITION OF VOLCANIC CINDERS Volcanic cinders are pyroclastic materials associated with recent volcanic activity. They occur in characteristically straight- sided cone-shaped hills which frequently have large concave depressions in their tops or sides where mixtures of solids and gases were released during the formation of the cone. Cinders vary in colour often within the same cone and may be red, brown, grey or black. The cinder particles also vary in size from large irregularly shaped lumps 50 cm in size, to sand and silt sizes. In some cones, however, particles may be more uniform with the largest size not exceeding 3 cm in diameter. Other characteristic features of cinders are their light weight, their rough vesicular surface and their high porosity. Usually they. are weak enough to be crushed under the heel. An advantage of cinders as a road construction material is the relative ease with which they can be dug from the quarry; a mechanical shovel or hand tools are usually adequate for their extraction although occasionally a bulldozer may be required to open up a working face. 3 FIELD SURVEY Field visits in connection with the survey were all carried out within a distance 150 km of Addis Ababa. They were concen- trated in areas near to Debrezeit, Nazaret, Zwai, Butajira and Giyon (see Fig 1). Approximately 70 cones were visited and more than 90 samples were collected for laboratory examination. Samples were obtained either from existing borrow pits from which material had previously been extracted or by digging pits where cinder cones had not been dis- turbed. Existing borrow pits provided the opportunity to obtain deeper profile samples which were more representative of the cone as a whole. 22 Fig. 1 Cinder cone distribution in Ethiopia, survey area enlarged Cinder cones rarely support any vegeta- tion other than grasses and examination of the exposed profiles showed that the depth of soil cover was not more than a few centimetres. A weathered cinder zone, however, usually extended down to a depth of about two metres. In some cones deposits of calcium carbonate coated the cinder gravels; these did not persist throughout the cones but in thin white bands parallel to and usually close to the surface. The size of cinder cones vary but they do not normally exceed 100 metres in height and side-slopes are generally of the order of 200 - 230. The largest cones would be expected to contain cinder deposits of about one million cubic metres. Occasionally cones occur singly but more commonly they are found in clusters in a linear arrangement associated with geo- logical faults and recent lava flows. The distinctive shape of the cinder cones made them easily identifiable on aerial photographs and photographs were used both to plan the survey and subse- quently in the field work. The examination of air-photos and print lay-downs was extended to cover the whole of Ethiopia and from these and a study of areas of 23 recent (Quaternary) volcanics on the geo- logical map, a preliminary map was compiled showing the occurrence of cones throughout the country (see Fig 1). They were mostly concentrated in the Rift Valley which extends from Tanzania and Kenya and bisects the country in a SSW-NNE direction; an indication of~ their frequency for each of the areas that were identified has been given. During the survey data sheets were used to make a comprehensive record of all the relevant information that related to each site. This included descriptions of soil profiles and samples that were collected, the geology, topography, vegetation and climate of the area. The data together with the subsequent test information that was obtained was used as an input to a data storage system for engineering materials which was developed as another part of the ERA/TRRIL Joint Road Research Pro~jectl. It is hoped that the data storage and retrieval system, when fully established, will be a valuable aid to highway engineers concerned with road planning and construction in Ethiopia. 4 LABORATORY INVESTIGATION Observations made during the field survey indicated that the two most important factors likely to affect the engineering behaviour of the cinder gravels were the grading and the strength of the gravel particles. The aims of the laboratory investigation on the cinder gravels, therefore, were to determine: i) the most appropriate methods for classifying and evaluating the engineering properties. ii) f rom same the variability of cinder gravels different deposits and within the deposit. iii) their strengths and densities at different levels of compaction. iv) the effect of compaction on the grading. v) the effect of moisture on engineering properties, and vi) the effect of adding locally available volcanic ash to make up for the deficiency of fine material (< 75 urn diameter) in the grading of cinders. The test methods used in the materials investigation were generally those given in the appropriate British Standard 2,3. 4.1 Classification tests 4.1.1 Particle-size distribution. Part- icle-size analyses were carried out on 53 of the samples that were collected during the field survey. Although the zone of weathering varied in different profiles it was clear that this did not extend beyond depths of two metres as, below this level, there was little fine material present. Out of 40 samples from below two metres that were tested only one, from a cone near to km 130 on the Nazareth to As~sela road, had more than 4 per cent passing the 75 pim sieve. The thirteen samples tested from the top two metres had a far wider range in the finer fraction with as much as 30 per cent passing the 75 pim sieve in a number of the profiles examined (see Fig 2). In some cases the fine fraction was plastic. - -- Samples from depths of less than 2m - - Samples from depths greater than 2m ~~Recommended grading for bases maximum size 38mm (1 'Ami) CL c0 0. 75 150 425 1.18 2.36 5 1 0 20 50 (pim) BS sieve size (mnm) Fig. 2 Gradings of samples obtained during the survey compared with a recommended grading envelope For practical purposes, it was the material below the weathered zone that was representative of cinder cones as a whole and although there was much similar- ity in the small amounts of fine material that they contained there was considerable variation in the size of the coarser material. The maximum size in some cones was no more than 2-3 cm whereas in others lumps up to 30 cm or 50 cm were present. Oversize material could be removed by screening or by crushing to produce smaller sizes but even so because of the deficiency in the fraction less the 75 pim the gradings did not conform to the recom- mended limits for road base materials 4 (see Fig 2). From the results of particle-size distribution it is clear that in order to obtain representative samples of cinder cones the material should be taken from below the weathered zone. 4.1.2 Modified aggregate impact test. The strength of the cinder gravel particles was measured by the aggregate impact test using the modified procedure form weak aggregates 5. The modified procedure is essentially a reduction in the standard number of blows of the falling weight that is applied to the sample and so avoids the problem of 'over-crushing' and compacting samples when weak materials are tested. The test proved to be very suitable for the work on cinders with a number of advantages over the Los Angeles abrasion test which is more commonly used in materials laboratories. The apparatus was portable and hand-operated and used relatively small quantities (about 300 g.) of material which could be tested in both the dry and soaked condition. A further advantage was the shorter time taken to complete the test. The modified aggregate impact values of 23 of the samples collected during the survey ranged from 46 to 177 (values expressed as that percentage of the original material which passes the BS No 2.36 sieve at the end of the test) and showed the wide range of strengths of the cinder particles (see Table 1). Even so they were all rated as weak when compared with the recommended values of less than 35 for aggregates used in the wearing course of bituminous road surfacings. Nine of the samples were also tested after a period of 48 hours immersed in water. The results showed that despite their porous nature there was no loss of strength except for one sample. From the description of the cinder gravel samples given in Table 1 it could also be seen that there was no correlation between colour and strength. This differs from the study by Hendrickson and Lund 6 on cinder gravels in the United States of America who found that dark coloured cinders (purple, grey and black) were harder than the lighter coloured red and brown cinders. 4.2 Variation of cinder gravel within a cone At three of the cones where material had previously been excavated, samples were collected to examine the variation of cinder within a cone. Particle-size distribution and modified 24 Table 1. Modified aggregate impact values of cinder gravels from different cones. Locaion(area) Colour of Modified aggregate Modified aggregate Location cinder gravel impact value impact value-soaked Bataj5ira Red 4*6 Lake Zwai. Red 50 53 Butajira Red 5 Butajira Red 58 Debre Zeit Red 59 Lake Zwai Light Brown 60 56 Butajira Black 67 Butajira Black 68 Akaki Red-browni 83 Debre Zeit Black 83 81 Bekojo Black 93 Bekojo Red 98 95 Lake Zwai Grey 100 81 Butajira Black 106 Butajira Red-black 107 Nazaret Black 113 110 Loggia Red 113 Nazaret Brown 126 N a zar et Red 128 Lake Zwai Light grey 131 162 Lake Zwai Red-brown 132 126 Modjo Black 14*5 130 Bekojo Black 177 Notes: 1. The modified aggregate impact values were determined on the 1~4 mm - 10 mm sized fraction. 2. Aggregate impact values of less than 35 are normally recom- mended for road surfacing aggregates. aggregate impact tests were carried out and the results are shown in Tables 2 and 3. The grading envelope for six samples from one of the cones, at Modjo, was similar to that obtained for all of the samples collected during the survey (cf Fig 3) illustrating that at depths below the weathered zone the amount of coarse material could vary but that the amount of fine material in the fraction less than 75 microns was consistently low. The modified impact values measured on a number of samples from each of three cones (at Modjo, at Bekojo and at Guse near to Butajira) showed that the strengths of cinder particles could also vary within the same cone. As the particle size increased then generally the strength increased,. 1uu[ 1 F 1 1 1 C,CS Cm,cl 80 - 60 - 40 . 0 75 150 '425 (pim) 1.18 2.36 5 1 0 20 50 BS sieve size (mm) Fig. 3 Grading envelope of samples from Modjo quarry 25 '.,00 Table 2. Modified aggregate impact values of same-sized samples obtained from different points in the same cinder core. Location ~Colour of Modified aggregate Location ~cinder gravel impact value Sample 1 Black 1142 " 2 Black 1414 3 Grey/black 129 Modjo " 4 Grey/black 121 5 Brown/red 1314 6 Red/brown 131 Sample 1 Red 514 Guse quarry ~' 2 Black 68 near to " 3 Brown 96 Butajira " 4 Black 66 5 Red 82 Note: Size of samples tested 114 mm - 10 mm. Table 3. Modified aggregate impact values of different sized samples from the same cinder cone. Locaticn Sie of sample Modified aggregate Locaticn Size of sample impact value Sample 1 50 mm 70 Modjo 2 .50 -25mm 83 3 25 -l14mm 110 " 4 114 -10omm 1145 Samplel1 50 -3 8mm 89 Bekojo. 2 38 - 25 mm 95 " 3 25 -l14mm 102 " 4 114-lo m 102 Note: The samples were crushed to 114 - 10 mm for the test. 14.3 Compaction and California Bearing Ratio (CBR) tests The compaction and CBR tests were important in obtaining further information on the engineering properties of the cinder gravels and the following factors were investigated. i) the effect of three different levels of compaction on the density/CBR/moisture content relation. ii) the effect of the different levels of compaction on the grading and the resultant changes in density and strength (CBR) that occurred. iii) the effect on the de'nsity and strength of soaking in water. iv) the effect on the strength of the addition of fines to make up for the deficiency of material in the less than 75 pim fraction. The cinder gravels used for the laboratory compaction and CBR tests were from cones at Modj5o and close to km 6, Debre Zeit - Bekojo road. Material from the cones had previously been used in nearby low-cost road projects and it was the DebreZeit - Bekojo road that was used to measure the road performance of cinder gravel which is described later in the paper. Materials from these two cones were also subsequently used in the full-scale experiment. The three levels of compaction 2 used in the tests were: i) British Standard 2.5 kg (5.5 lb) hammer method. ii), British Standard 14.5 kg (10 lb) hammer method. iii) British Standard Vibrating hammer method. Because compaction was expected to cause breakdown of the weak aggregate particles two series of tests were carried out 26. initially which both involved measuring the changes in grading. In the first series separate samples were used for each point in the density/CBR/moisture content relation with particle-size analyses deter- mined after each point, and in the second series the same material was used for all of the points of the dry density/CBR/ moisture content relation. In the second series, therefore, the materials was sub- .jected to repeated cycles of compaction and particle size analyses were determined after the first, third and fifth cycle. In both series of tests the results obtained for the two cinder gravels followed a similar pattern. For illustra- tive pruposes, therefore, only one of the samples, that from Mvod~jo is referred to in detail. These results, given in Figs 4 to 9, showed that: 0 A c W U 1600 1400 I- 1200 [- .0 c0W 1000 [-- 800 L i) Higher densities and CBR values were obtained as the compactive effort increased but,unusually, in the first series of tests using separate samples there was virtually no change in dry density or CBR values as the moisture content increased (Fig 4). ii) In the second series of tests the dry density and CBR values did increase with moisture content (Fig 6). o BS compaction (4.5 kg), BS8 vibratory compaction ABS compaction (2.5 kg) cc C-) 85 heavy compaction (4.5 kg) BS vibratory compaction as compaction (2.5 kg) 80 0 60 0 0 40 -0-- -0 -- 20 A A A AA 90 ~ 0 0~ 0 80~~~~~~~~ A A A A A 60 0 5 10 1 5 Moisture content (per cent) 20 Fig. 4 Dry density/CBR/ Moisture content at different levels of compaction. Separate samples used for each point on the compaction curve 1600 r- 1400 1- 0 0 200 F- .0 *0 W 1 000 l-* 800 L. 0 5 10 1 5 20 25 Moisture content (per cent) Fig. 6 Dry density/CBR/Moisture content relationship at three levels of compaction. The same sample used at each level of compaction iii) The difference in the results of th~e two series of tests can be explained by the effect of compaction of the grading of the materials. The first series showed that different levels of compaction produced different amounts of breakdown (Fig 5). 27 c n 1 ~~Original grading - --6E5 compaction (2.5 kg) L- .- BS compaction (4.5 kg) ....... 85 vibratory compaction C) cl 0~ ~~Original grading - .- One cycle of compaction - Three cycles of compaction -- Five cycles of compaction 75 150 425 1.18 2.36 5 1 0 20 (pm) BS sieve size (mnm) Fig. 5 Effect of compaction on grading-Separate samples used for each point. Each curve is an average of five points In the second series, however, because the seine sample was used, there was an additional, cumulative effect of com- paction which is shown by the increased amount of breakdown between the first, third and fifth cycles (Figs 7, 8 and 9). The difference between the third and fifth cycle though was only marginal and showed that the changes do become less as maximum packing of the particles by com- paction is achieved. Original grading - .- One cycle of compaction - - Three cycles of compaction -.- Five cycles of comnpaction 75 150 425 1.18 2.36 5 1 0 20 (pmr) BS sieve size (mm) Fig. 8 Effect of repeated compaction on grading, 85 vibrating compaction ~~Original grading - One cycle of compaction - - Three cycles of compaction -- Five cycles of compaction 100 80 m LCu 60 cm40CU 20 75 150 425 1.18 2.36 5 1 0 20 (mm) BS sieve size (mmn) Fig. 9 Effect of repeated compaction on grading, BS compaction (4.5 kg) _____ ___ ~~~~iv) The effect of compa'ction resulted in - _____ ___ ___ ___ ~improved gradings for the cinder gravels ___ ___ ____ ___ ~~which, apart from a deficiency of fine material in the fraction less than 75 - ~~~- ~microns, now conformed to the recommended 75 150 425 1.18 2.36 5 1 0 20 limits for materials used in road bases. (pmr) 65 sieve size (mmn) v) The increase in density and CBR values, because of the change of grading, was Fig. 7 Ef fect of repeated compaction on grading, appreciable with maximum CBR values of 75 BS compaction (2.5 kg) per cent and 105 per cent being obtained 28 100 CL CD0. C,W CL for the two cinder gravels. These results indicated that if the same levels of dens- ity were reached in practice in road con- struction the cinder gravels were likely to prove suitable for use in road bases. vi) CBR tests on the cinder gravels after being immersed in water for four days following compaction showed that there was no loss of strength. 4.4 Mechanical stabilization: the addi- tion of fines to cinder gravel. It has been shown that the effect of compac tion improved the grading of cinder gravels but that there was still a deficiency in the fine fraction of material less than 75 microns. A further series of tests was, therefore, carried out with additional fine material added. In practice this would appear to be a feasible proposition as suitable material, volcanic ash, is invariably found adjacent to cinder cones and so is available locally In the laboratory investigation, the Modjo cinder gravel was used and 10 per cent by weight of clayey ash soil from near the cone was added. The soil had a liquid limit of 64 per cent and a plas- ticity index of 24. The tests carried out were the determin- ation of dry density/CBR/moisture content relations at two levels of compaction (2.5 kg and 4.5 kg hammers) with separate' samples used for each point of-the test as in the first series of compaction tests. The effect of immersion in water for four days was also examined. The results are given in Table 4 and in Fig 10 which compares the density/CBR/ moisture content relations for the cinder gravel with and without the fines added. The addition of fines increased the dry density for both levels of compaction by about 8 per cent and the CBR values were also increased with an improvement from 60 per cent to 80 per cent for the heavier level of compaction and from 18 per cent to about 26 per cent for the lighter compaction. Soaking in water for 4 days reduced these values to those for the cinders without fines added. A further effect of the addition of fines was that the moisture content at compaction could influence the density and CBR obtained, although not in a very marked way. 5. INVESTIGATION OF A CINDER GRAVEL ROAD It has been stated that in order to pro- vide recommendations for the use of cinder Compaction level 2.5 kg 4.5 kg A A Without fines 0 * With 1 0% fines added 1600 - 1400 ;Z 0 1200 I- 1000 800 L 0 5 10 1 5 20 25 30 Moisture content (per cent) Fig. 10 Comparison of density/CBR/Moisture content relation for Modjo cinder gravel with and without fines added gravels, an understanding of their engineer- ing properties and performance in the road was required. A gravel road about 20 kilometres long between the town of Debre- Zeit and the small village of Bekojo, which had been in service for about two years and was constructed with cinder gravel taken from a cinder cone close to kilometre 6 from DebreZeit, provided the opportunity to carry out an investigation to obtain data on the performance of the material used, and to make a comparison with the results of the laboratory investigation. The road, which was first visited at the end of the rainy season and was in good condition with only slight rutting in the wheel-tracks and few signs of corrugations had been built by labour-intensive methods. Construction plant had been confined to the use of a blade-grader to spread the material across the formation width of 6 metres, with compaction being left to the 29 Table 4. Dry density fines added. and CBR values for Modjo cinder gravel with 10 per cent of Level of compaction k/3Dydensity lbf3 Moisture content CBR lb/ft3 - per cent - per cent 1074 67.1 6.3 25 British tandard1102 68.8 11.7 29 Britis hSandard 1105 69.0 17.2 26 2.5 kg hammer 1116 69.7 21.8 25 1o84 67.7 26.3 18 4 days soaking 1073 67.0 5.0 ~Before 19 1119 69.9 18.0 soaking 20 1345 84.o 6.4 66 British tandard1400 87.4 l0.6 8 Britis hSandard 1427 89.1 15.6 79 4.5 kg hammer 1435 89.6 20.3 66 1440 89.9 23.6 56 4 days soaking 1341 83.7 5.0 ~Before 60 11 11 11 ~~1428 89.2 18.0 ~soaking 55 action of local traffic which was of the order of 20 vehicles per day. 5.1 In-situ density tests and particle- size analyses. Three sites were chosen at intervals along the road to measure the in-situ density and grading of the cinder gravel. At each site measurements were made at five points across the road with one on the centre-line, two in the outside wheel- tracks and two in the inside wheel-tracks where the influence of traffic on the compacted density and on the grading was expected to be greatest. The densities were determined by the sand replacement method 2 and samples for grading analysis were taken from positions ad~jacent to the density holes. Results of the road investigation and comparisons with some of the results of the laboratory investigation are given in Table 5 and Figs 11 and 12. Additional measurements were made at each site of surface deformation, the thickness of the gravel layer as well as the moisture con- tent and plasticity of~ the subgrade soil but these results are not reported here. The results of the in-situ density and grading tests at each of the three sites on the road showed that: i) The highest densities and greatest degree of breakdown occurred in the position of the wheel-tracks (Table 5 and Fig 12). ii) The mean value for the in-situ densi- ties in the wheel-tracks was 1538 kg/rn 3 (96 lb/ft 3). This was higher than those obtained for both compaction levels (2.5 kg hammer and 4.5 kg hammer) using separate samples and the same as that for L- - - Inside wheel tracks --- Outside wheel tracks .* .. ..Centre line -~ Original grading from quarry face 100 01CL 0)m CL 1 75 150 425 1.18 2.36 5 1 0 20 50 .(M m) BS sieve size (mm) Fig. 1 1 Grading of samples taken at km 2 Debre Zeit-Bekojo Rd ~~19mm grading for base - -- Laboratory grading after 5 cycles of 85 compaction (4.5 kg) --- - Grading from wheel track Bekojo Road 0. 0) CL 40 _ _ _ _ _ _ _ _ 20 ____ - 75 150 (pm) 425 1.18 2.36 5 10 20 BS sieve size (mm) Fig. 12 Comparison of laboratory and in-situ grading with the recommended 19mm M~n) grading envelope for bases 30 Ln o LA LA Ln CN C') C') C') C') w. tLA Ln LAnw. O (N NI N1 ND 03 C4iC C') 013 U) hd a) 3-I H) U) t3U) CD C CD l No N W 03 W' m m cm LAC' H H cm o. o. LA LA LA C') ND mn um 03 .~r ~1 .;0C'4N0'M -~ H H H H N- N- 0" N '0 0303030303 C "'T1 ~ .0 C) 03D O') 03' C') m r (n M , m rn( H4 -H H H4 H "' en 'I LA LA m~ 4 C') m~ '. "3.P rn '0 W W N O,1 H m~ 0" CD O '0 0303 M R m" '.0 M m 0 .;5 '.0 LA LA in H4 H- HA H- H o oLn03 CON Ln0 0 ,-4 W r~ N~ 1 m W O C') 0 3N LA H W. m 03 H4 m 0-~4 '0 4-i 5-i 0 U) tp d. .0 U) Q m .n n a) C -iD C COw m M m, MM' 03 04-3 4J 0 4J C:~ NV0 0 H H H H H . (n5) 31 4-i0.a)X 04 a) 4-iU).-40 4-i0-a)4-i0~0U 4)30~a1) a1)cl m4)34-4 .0H m0=tr .14 (L) 0 0 ----p > 04 A ~0U '0 CO10it4.3U) .0~U)-144)i-4s-I U) a) ( 31 W P4LAW c0 4-iid -'-44-iU) 0)'-I Hd B- 0_0-144)CO -44) H '0 0 00-44-iCd "-44)U)a) 0-4 04-i .1H 0-4 0a)0-4Cd4-i.00C))a)-44-i-4U)0a)'0 .04-i-4 U)a)-44-i-4La0a)'0 :34-i-4LI) 0-4'4-40 00U)'-45-ICd0400C-) LA a,'-I 4i 0) 0 a) 4JU)-140 4)3 4)0:00 m4)3 >1 4-1 4) 0)r.0)m .54 0:04)-14U)00'4 Q) U)Q CO 4) a-1a) a)4i 0 .54 34J-4H 3-4a)0:0lH -0 04 3- 4 3-I W WH H : H 0 _. 4.54 .54 0 0 0 Cd (U Cd 4) 4) 4) a) a) -N aW .C.0H.0 1 ~: a) 5-43-13-Il4a)W a4) a 4i r c:0. : 0 a )W 0 0 H UH c) I-4H 4)0.a)U .540Cd3-44) H1a)a).0 '-Ia)4 ):30 H- C') the heavy compaction when repeated com- paction of' the same sample was used. At that density the laboratory CBR value was 80 per cent (see Table 5). iii) Trafficking produced changes in the grading of' the cinder gravel similar to those obtained in the laboratory com- paction tests. The improvement was sufficient in the position of the wheel- tracks, to meet the normally recommended limits for road base materials. 6. CONCLUSIONS The main conclusions from this first stage of the investigation of cinder gravels which comprised a field survey, a labora- tory study and an examination of a cinder gravel road are: i) Cinder gravels are more widespread in Ethiopia than was originally believed. This showed the value of using aerial photographs in survey work and enabled a preliminary map to be prepared giving the distribution of cinder cones. ii) In o:-der to obtain representative material from a cinder cone it is impor- tant that samples are taken from below the weathered zone which can extend to a depth of two metres.. iii) Although "as-dug" cinder gravels do not meet the recommended grading require- ments for road base materials the labora- tory investigation revealed that, because of the weak nature of the aggregate particles, breakdown under compaction occurred with an improvement in both grading and strength properties. iv) In the laboratory investigation the cinder gravels 'were not affected by changes in moisture and even complete immersion in water only reduced their strength slightly. v) The addition of locally available clayey soil to make up for the deficiency of fine material in the grading improved the stability of cinder gravels and indicated that this could be a useful construction practice. vi) The gravel road study confirmed that an improvement in the grading and the strength of cinder gravels occurred under normal road conditions, even when the means of compaction was simply the effect of traffic. The indications from the study carried out so far are that cinder gravels should make useful road construction materials especially for gravel roads. Further work, however, is necessary to examine a range of cinder gravels'under known conditions of traffic and climate in bituminous surfaced roads as well as in gravel roads before limits can be recom- mended for their various uses. This was the purpose of the full-scale road experiment which is in progress. 7. ACKNOWLEDGEMENT The work was carried out as part of the programme of the Joint Road Research Project of the Ethiopian Roads Authority and the Overseas Unit of the Transport and Road Research Laboratory, United Kingdom, and the paper is published by permission of the Director. The authors wish to thank members of the Materials and Research Branch, ERA, and members of the Overseas Unit team in Ethiopia who gave so much help to the study. 8. REFERENCES 1. Lawrance, C.J. The use of punched cards in the storage and retrieval of engineering information. ERA/TRRL (UK) Joint Road Research Project JRRP Report No 2 Addis Ababa 1975 (Ethiopian Road Authority). 2. British Standards Institution. British Standard 1377:1975. Methods of test for soils for civil engin'eering purposes. London, 1975 (British Standards Institution). 3. British Standards Institution. British Standard 812:1975. Methods for sampling and testing of mineral aggre- gates sands and fillers. London, 1975 (British Standards Institution). 14. Road Research Laboratory. Soil mechanics for road engineers. Depart- ment of Scientific and Industrial Research. London, 1952 (HM Stationery Office). 5. Hosking, J.R., and Tubey, L.W. Research on low-grade and unsound aggregates. Ministry of Transport, RRL Report 293. Crowthorne, 1969 (Road Research Laboratory). 6. Hendrickson, L.G., and Lund, J.W. Construction specifications for volcanic cinders used as road surfacing aggregate. Highway Research Record No 307 1970. Crown Copyright 1979. Any views expressed in this Paper are not necessarily those of the Department of the Environment or of the Department of Transport. Extracts from the text may be reproduced, except for commercial purposes provided the source is acknowledged. Reproduced by permission the controller of Her Britannic. Majesty's Stationery Office. 32