Letters to the Editor

Letters to the Editor Sir : THE MOHENJO-DARO FLOODS: A REJOINDER! Mr. Possehl s article The Mohenjo-daro Floods: A Reply (1967) both deserves and requires a strong rejoinder. The persons directly referred to in his Reply feel that he has seriously misread and misinterpreted the evidence, which has been considerahly augmented since he wrote (see appended bibliography). It has been disappointing to read such statements as the proposers of the theory did not attempt... and the investigators tell us pitifully little.... Mr. Possehl knows the difficult circumstances surrounding these investigations and that they were suspended indefinitely in 1965. As a general comment it should also be pointed out that journals such as the Anthropologist are not the place for the presentation of raw data concerning the levels mentioned, the soils tests carried out on borehole samples, and so on. We would welcome nothing more than the opportunity to pursue the field research and to publish the data and results in full. But until that time we feel comfortable that what new evidence there is has been presented in an objective manner and that an important new dimension has been added to the debate. The first part of this rejoinder, by Mr. Raikes, deals with the technical aspects of the flood theory. Dr. Dales, Director of the 1964/65 Mohenjo-daro project, follows with some comments on the archeological aspects. Raikes Rejoinder The first point to clarify is that the theory postulated by this writer grew by stages. The original tentative theory had its origins in about 1961 and was eventually published in this journal in 1964. It had its origins in observations, supported by much already published geological evidence in Baluchistan and in the anomaly here that cannot be simply explained in terms of abnormal floods... referred to in The End of the Ancient Cities of the Indus (Raikes 1964:289). At that time no new excavation had been carried out at Mohenjo-daro and none of us had much reason to believe there were more than 30 feet of buried deposits. Nothing substantially below the water table, at some 12 to 15 feet below the plain level, had been scientifically excavated because of the then insuperable difliculty posed by the water. In that first paper I did not go beyond the suggestion that the dating evidence might require reviewing. The dating evidence in fact was very scanty and highly unreliable, as Dr. Dales will demonstrate. 957 There are two highly relevant points that concern that first paper and Mr. Possehl s reaction to it. The first concerns the wearing out of the landscape, quoted uncritically from Wheeler (1966: 77). In this context, presumably, wearing out the landscape is the same as exhausting the soil. One cannot have it both ways: if the Indus Valley was being subjected to abnormal flowing floods, or to inundation of whatever origin, such as to raise the level of the plain by some 70 feet in (to take Possehl s figure, 1967:39) 500 years, each year saw the addition of an average depth of fresh sediments of between 14 and 2 inches. It would certainly be dithcult to exhaust the soil s resources under such circumstances; indeed with relatively primitive and unintensive cropping it would be virtually impossible. It is difficult to believe that abandonment of the peasant settlements in Baluchistan was due to exhaustion of the soil, seeing that nearly all cultivation is carried out on silt-trapping terraces. Elsewhere (1967:38) Mr. Possehl refers to extensive soil erosion. Soil erosion and a steady build-up of sediments?! The second point concerns the assertion of which I am accused by Mr. Possehl (1967:39) concerning the abandonment of the Baluch villages by 2300 B.C. It was not an assertion. What I wrote was On the whole it seems to fit in much better with archaeological evidence and that of geology to hypothesize that these peasant cultures in the hills came to a fairly sudden end due to the same causes that destroyed Mohenjo-daro and that a fairly early date, perhaps even as early as 2300 B.c., should be considered for this period of tectonic activity (1964: 295) (italics added). Mr. Possehl s own argument for contemporaneous existence of the Harappan cultures and those in Baluchistan does not of itself invalidate the possibility that all of them ended at around the same time, but that time has still to be determined. (However, see Dales s remarks below.) If my reference to this period of tectonic activity has been construed as covering only a few years, I have signally failed to communicate. The existence of multiple raised beaches along the Makran coast points to a series of movements, at unknown intervals of time, along the southern limb of the Baluchistan geosyncline and it could have taken several centuries for abandonment to have resulted, whether in Baluchistan or in the lower Indus Valley. The hill area is subject at all times to earthquakes, but these by themselves do no more than temporarily interfere with settled life. However, if a period of repeated episodes of major earth movements gradually changed (as the geological evidence shows) the topographical relationship between settlements and their sources of water, a moment could have been

958 American Anthropologist [70, 19681 reached for each settlement at which economic survival on a peasant economy became impossible (not hy any means necessarily the same moment everywhere). As at Mohenjo-daro, a fairly long period must he accepted for the gradual engulfment of the settlements in the area roughly between Jhukar and Sehwan. This area contained a very large concentration of Harappan sites. When Mr. I ossehl(1967:36) writes I just do not see that Mohenjo-daro was ever engulfed in mud, one is tempted to ask just what he does see. To this writer it seems entirely immaterial in this context whether the mud -actually silty clay-was deposited by flowing floods or under still water. Whatever the cause the inescapable fact remains that Mohenjo-daro and other sites in the area were most effectively engulfed in mud at least 40 feet of which still survive under the flood plain level; for this is no isolated jell formation, the whole level of the plain having risen by at least that amount and almost certainly by 70 feet or more. If this is not engulfment by mud we are in danger of becoming engulfed in semantics. If Mr. Possehl can postulate another credible explanation for the deposition of 40 (or 70) feet of silt above the original flood plain in a few hundred years, and then no further deposition, all those interested in unraveling the story of Mohenjo-daro will be grateful. (Let us be spared, please, any resort to arguments based on deglaciation around ZOO0 B.c.!) For thcre can be little doubt that the silty sand, of very great depth, with its thin cover of a yellowish clay, revealed by borings (not merely those at Mohenjo-daro) represents the flood plain at the time of the founding of Mohenjo-daro. Possehl s explanation should also account for the localization of this gradual engulfment by mud in only one part of the vast and otherwise remarkably uniform length of the Indus Valley. Conclusive proof of the existence of the postulated dam in the form of terraces surviving the rejuvenation of the river was of course sought for. But the passing of at least 3500 years and the known recent history of channel changes in the Indus-not to mention the colossal changes wrought by man in creating one of the world s largest irrigation systems-make terrace survival unlikely. Terraces do in fact appear to exist, with alignment at least not inconsistent with what one might expect to find from lake-fill terrace remnants. But any attempt to identify them as such or to distinguish them from normal terracing of the Indus (whatever that may mean to others, it means little or nothing to me in the context that follows) in the light of its known recent history would have been irresponsible. The suggestion that the proposers of the theory did not attempt to find them (terraces) is not very flattering and is in fact untrue. But if Mr. Possehl is to distinguish, either from air photos or on the ground, between ruined sailuba terraces on the relatively steep flanks of the valley and lake-fill terrace remnants, he will have to devote many months of patient search on the ground and probably finish up no wiser than when he started. The case of Lake Manchar, another of Possehl s points, is extremely relevant. The lake stands at present at a somewhat higher altitude than the Indus river at Sehwan. In the lake itself are Hardppan sites almost totally engulfed in the silt deposits of the lake. Whether the sites around Lake Manchar mentioned by Mr. Possehl should now be completely hidden from view depends on a number of factors; those at Pandi Wahi and Tando Rahim Khan being on lateral coarse alluvial fans at the debouchment of Baluchistan torrents were probably never affected by the dam. In the absence of complete excavation at Lohri, Lakhiyo, and Gazi Shah, the extent to which they are buried is unknown, but it seems highly probable that they were completely buried before rejuvenation 01 the valley. In fact Lake Manchar may well be on a terrace remnant from some intermediate stage of rejuvenation. I cannot see why it should be any more strange for there to be Harappan and even pre-harappan sites in this small region than that a similar site should have existed at Lohumjo-daro. A point that was overlooked in my previous papers, perhaps because it was so evident to me as to seem hardly worthy of note, is the absence of evidence of damage at Mohenjo-daro caused by flowing water. Normal flooding of the Indus with fast moving waters, suggested by Possehl as responsible for the Mohenjo-daro silts, would necessarily have left evidence of extensive damage. I do not know what modem map has suggested to Mr. Possehl that a dam of 150 miles length would have been required. My maps indicate that a length of some 50 miles (about the same as Allah Bund) would have sufficed. I find it hard to believe that an elementary confusion between feet and meters has caused his exaggerated estimate. Indeed the investigators said pitifully little about the silt levels. The reason should be obvious. The reports so far made have been interim reports in article form. No final report has been made because the work has been suspended. Articles in Antiquity or this journal are not generally considered the right place for the publication of raw data such as the levels mentioned or the soils tests carried out on bore-hole samples. The criteria used to distinguish the origin of the deposits were grading analyses and Atterberg limits tests. The former show percentages of clay that center around 15 percent, if the MTI classification is used, or around 40 percent, if the clay fraction is considered as all material of 0.005 mm grain size and under. The latter show that all samples, except one identified visually as sandy clayey silt, lie near the borderline between Inorganic Clays of Low Plasticity and Inorganic Clays of Medium Plasticity. This material is, as regards its clay content, depositable only in still water.

When clay is deposited, ipsofuc/o all coarser material is also deposited. Even the sandy clayey silt, identified on site and afterwards analyzed in a laboratory, contained 15 percent clay on MTI classification or 22 percent on the older classification, and it came within the same classification based on Atterberg limits tests. To avoid any possible misunderstanding it must be emphasized that two of the samples tested were from near the top of the new excavations and showed analyses indistinguishable from those of samples taken from depths varying from 11 0 to 36 0 below ground. Of the numerous samples six were selected for detailed analysis, with the intention of extending the analysis (and cost) only if the results showed inconsistencies. In fact the results were so extraordinarily consistent that there appeared to be no justification for incurring additional expenditure. Field tests are not infallible, even when carried out by an experienced person, but within the normal limits of experienced human error I can say that the samples examined from other parts of the large site were indistinguishable from those fully tested. The reference to the possibility of the formation of oxbow lakes is really too far-fetched to be taken seriously. As however it may be taken seriously, let us consider what it would involve. In order to account for the continuous deposition of still-water deposits over the whole area of the city and to such a great depth, the oxbow lake would have to be of quite exceptional size and would have to remain stationary throughout the whole up-building period. Meanders and their oxbows are by definition mobile phenomena; need one say more? Yes, perhaps it should be added that definition of the Indus as of mature character today surely is no proof that it was similarly mature when either 30 feet higher than today or 40 feet lower. As regards those levels there is a misprint on p. 36 at the foot of the left-hand column. This should read: There are only three zones, between +155.5 and +158.5, +168.5 and +170, and +175.2 and +176.7... (1967). The italics added by Mr. Possehl at the head of the same page ( that seemed to correspond ) are gratefully acknowledged, for emphasis of that verb was certainly intended and is of the greatest relevance. At the time when the survey was carried out the dam theory had reached a stage (later found untenable) at which it seemed reasonable to look for general inundation levels. It had not then been appreciated that such levels could only have been associated with depths of water that would have drowned (not engulfed) the city. The random distribution of levels at Mohenjodaro was one of the factors that led to storage calculations that in their turn involved the absolute necessity of an approximate equilibrium between inflow on the one hand and leakage and evaporation losses on the other hand. This is where the sand dam, whose truncated base has been identified at Letters to the Editor 959 two points near Sehwan, came into the picture. Of the 150 points leveled, 70 were at the apparent top surface of deep sterile silty day (apparent because in a few cases, at least, there was some evidence of cutting of foundations into this material); another 52 were on top of rubble saturated with silty clay; and 28 were cases including identifiable mud-brick or otherwise doubtful deposits. All cases, if one excludes the two or three of identifiable mud-brick, represent levels at which clay occurred with an appreciable thickness (up to a maximum of 7 feet or so) below the point leveled. The gap between +155.5 and +158.5 is almost certainly due to the dsculty of finding any identifiable samples near existing ground level, owing to taluses of collapsed material. The other smaller gaps from +168.5 to +170, and from +175.2 to +176.7 represent intervals only between tops of samples leveled. In fact, silty clay deposits extend downwards for a great deal more than the gap in each case. The case might have been better expressed, and I am glad of this opportunity of doing so, as follows: With the sole exception of the zone from +155.5 to 158.5 (which may only be an exception fortuitously), the whole height of the mounds from+158.5 to +186.5 has silty clay at every level. In other words, over the mounds as a whole there is evidence of continuous building up of silty clay deposits. If it were not for the anomaly that Mr. Possehl s argument ignores, this could be interpreted (though wrongly) as the result of regular over-spill floodingif it were not for its still-water character. Dales Rejoinder Mr. Possehl appears overanxious to reject totally the natural disaster theory and substitute yet another theory, apparently that of Dr. Fairservis, his mentor, that is built on highly subjective and controversial premises. At least the natural disaster theory is founded on observable scientific facts, incomplete as they admittedly are at present. A full appraisal should include comments on Fairservis recent monograph (1967), but it arrived too late to be included here. Factually, my first criticisms concern dating and chronology. This is crucial matter and one in which all of us at one time or another-including myselfhave exhibited more reliance on faith than on fact. There are three major considerations here: (1) the relative chronology of the Indus-Baluchistan cultures, (2) the absolute dating of the Indus civilization based on radiocarbon (C-14) dates, and (3) the correlation of Indus and Mesopotamian chronologies and historical events. A detailed discussion of (2) must wait for the results of a study I am making in collaboration with the Radiocarbon Laboratory of the University of Pennsylvania. Suffice it to note that Possehl s insistence on 2250 B.C. as the upper limit for Mature Harappan is untenable on present evidence. Either the generally

accepted date for the Mesopotamian Sargonid period-for which material contacts are substantiated between Mesopotamia and the Harappans (Dales 1968)-is a century too high or the Harappan chronology based on radiocarbon is too low. As for the lower limit for $Mature Harappan, here again there is uncertainty, partly from lack of sufficient archeological information and partly from careless, or untutored, use of the available radiocarbon dates. Analysis of the available dates, including careful attention to such factors as the pretreatment of the samples, their archeological contexts, calculation of mean-averages for related groups of samples, and the use of 5730 years as the half-life, suggests that LLmature Harappan ended closer to 1900 than to 1700 B.C. Possehl s discussion of the relative dating of Harappan civilization and the Baluchistan village cultures is an example of the faith rather than fact mentioned above. First, he commits a cardinal archeological sin when he quotes a single radiocarbon date as proof of the real age of anything (1967:39). It is now generally recognized that radiocarbon calculations are subject to too many variables to permit any high degree of credibility to single samples. This is especially pertinent here, where even the archeological context of this particular sample is questionable. Another serious misuse, and apparent misunderstanding, of the use of radiocarbon dates is reflected in Possehl s statement (1967:39) that Fairservis dates the end of the Baluchistan Damb Sadaat I11 phase as late as 1650 B.C. The samples involved were collected and tested during the early days of radiocarbon experimentation when such phenomena as the Seuss effect were not recognized. Once the effect was recognized, it was announced that a 200 year correction factor must be added to all of the early dating calculations. The samples Possehl refers to here (published originally in Fairservis Quetta report) do not include this correction. Also, these samples were tested by the old solid-carbon procedures, and they received no pretreatment for the elimination of possible humic acid contamination. Furthermore, the dates quoted require correction for the new antl widely accepted half-life of 5730 years. This adds approximately 120 years to the dates (making them older). Even given all these technical corrections the dates in question have such wide tolerances-as much asf 400 years!-that they are virtually meaningless. Neither these, more recent radiocarbon dates, nor the comparative archeological evidence suggest that this apparently terminal phase of occupation in the Quetta area of Baluchistan was later than the earliest phase of Harappan civilization (Dales 1965). The absolute dates for this as yet unidentified early Harappan phase are speculative but circumstantial evidence points to somewhere in the 2500 to 2300 B.C. range. Such careless use of dating evidence can be disas- American Anthropologist [70, 19681 trous to one of the basic aims of archeologists and historians, namely, to reconstruct the chronological development of a region and its population. The chronology of Baluchistan and the Indus Valley is still fraught with more questions than answers. Such dogmatic-and often erroneous-assertions as presented by Possehl (and by inference by Dr. Fairservis) can do little to further our understanding of these complicated problems. The chronological questions are directly relevant to the problem of Harappan decline. This artificial lowering of the entire chronological structure of Baluchistan by Possehl (1967: 39) seriously distorts the evidence, apparently to make it fit into a preconceived scheme of the progressive impoverishment of the Baluchistan and Indus Valley peoples. If one accepts Possehl s relative chronology, then perhaps one can feel comfortable with his impoverishment theory. On the other hand, Raikes antl I feel that it is a pure fabrication based on subjective ecological speculations and a distorted chronological framework. The flood theory for Mohenjo-daro and the central Indus region proposed by Kaikes and supported by me is hut part of a much larger and more significant picture of natural disasters that apparently confronted and defeated both the Baluchistan and Indus Valley populations. These other factors require extensive and intensive investigations but their potential significance cannot be disregarded or peremptorily rejected. For example, Raikes has described the geomorphological factors that might have had disastrous effects on the Baluchistan populations (rejected by Possehl, 1967:39, mainly on chronological grounds). Also there is the question of distribution of Harappan sites in the dried-up regions of Bahawalpur and northern Rajasthan. Geologists and geomorphologists with firsthand knowledge of the lower Himalayas have made the observation that natural disturbances in the headwater area could account for the diversion of waters from the Sarasvati-Indus system over to the Gangetic. There is also the situation of the Makran coast where geomorphological changes apparently altered the strategic positions of the Harappan seaports (Dales 1962). Practically all investigators of Harappan sites in the lower Indus and Gujarat regions have commented on the evidence for disastrous effects of flooding. Lastly, the recent investigations of geomorphologists (e.g., Snead) have provided impressive evidence substantiating the geological instabilities of Baluchistan-lower Indus regions. All of these lies of evidence, and others, point to the conclusion that natural disruption played an important role in the life and death of the Harappans. The socioeconomic and cultural factors that certainly were involved in the final demise were brought about, or at least intensified, by the natural disasters. Much more could and should be said concerning Possehl s interpretation of evidence (especially about

the origins and nature of Harappan civilization) but the main purpose of this rejoinder is to set straight some of the basic data relevant to the role of natural forces in the demise of Harappan civilization. In conclusion, the leaking dam theory proposed by Raikes is the only hypothesis that satisfies both the archeological and geomorphological conditions. Further interpretations must be based on such hard facts rather than upon preconceived socioeconomic models. GEORGE P. DALES University Museum Unioersily of Pennsylvania Letters to the Editor 961 ROBERT L. RAIKES Raikes and Partners Rome, Ilaly REFERENCES CITED DALES, GEORGE 1:. 1962 Harappan outposts on the Rlakran coast. Antiquity 36:56-92. 1965 A suggested chronology for Afghanistan, Baluchistan, and the Indus Valley. In Chronologies in Old World archaeology. Robert Ehrich, ed. Chicago, University of Chicago Press. Pp. 257-284. 1968 Of dice and men. Journal of the American Oriental Society 87 (in press). FAIRSERVIS, W. A. 1967 The origin, character, and decline of an early civilization. American Museum Novitates, No. 2302. POSSEHL, GREGORY L. 1967 The Mohenjo-daro floods: a reply. American Anthropologist 69:324. RAKES, ROBERT L. 1964 The end of the ancient cities of the Indus. American Anthropologist 66: 284-289. WHEELER, SIR M. 1966 Civilizations of the Indus valley and beyond. New York, McGraw-Hill. ADDITIONAL REFERENCES (see also those cited by Possehl) AGRAWAL, D. P. 1966 Harappan chronology: A re-examination of the evidence. In Studies in prehistory. D. Sen and A. K. Ghosh, eds. Calcutta, K. I,. Mukhopadhyay. Pp. 139-148. DALES, G. F. 1964 The mythical massacre at Mohenjo-daro. Expedition 6(3) :3643. 1965a Civilization and floods in the Indus Valley. Expedition 7(4) : 10-19. 1965b New investigations at Mohenjo-daro. Archaeology 18: 145-150. 196% Re-opening Mohenjo-daro excavations. Illustrated London News, May 29: 25-27. 1966 Recent trends in the pre- and protohistoric archaeology of South Asia. Proceedings of the American Philosophical Society 110: 130-139. FAIRSERVIS, W. A. 1966 Problems in post-harappan archaeology in the lower Indus Valley and Baluchistan. Jour- nal of the Maharaja Sayajirao University of Baroda 15(1):2.5-36. GHOSH, A. 1965 The Indus civilization: its origins, authors, extent and chronology. In Indian prehistory: 1964. V. N. Misra and M. S. Mate, eds. Poona, Deccan College. Pp. 113-156. LAL, B. B. 1963 A picture emerges: an assessment of the carbon-14 datings of protohistoric cultures of the Indo-Pakistan subcontinent. Ancient India 18 & 19:208-221. RAIICES, R. L. 1965 The Mohenjo-daro floods. Antiquity 39: 196203. 1967a The Mohenjo-daro floods-further notes. Antiquity 41 : 64-66. 1967b Water, weather and prehistory. London, John Baker. SNEAD, R. E. 1964 Active mud volcanoes of Baluchistan, West Pakistan. The Geographical Review 54: 546-560. 1966 Physical geography reconnaissance: Las Bela coastal plain, West Pakistan. Indian Ocean Studies, Technical Report 15(1). Baton Rouge, Coastal Studies Institute, Louisiana State University. STUCJCENRATH, R., JR. 1967 University of Pennsylvania radiocarbon dates X. Radiocarbon 9:333-345. Sir : FURT- COMMENT ON THE AUSTRALIAN SUBINCISION CEREMONY Correspondents in the USA. and Australia have written to me drawing attention to The Australian Subincision Ceremony Reconsidered: Vaginal Envy or Kangaroo Bifid Penis Envy (AA 69: 355-358). They have expressed surprise that the authors, Drs. Singer and De- Sole, overlooked recent fieldwork available in an international psychiatric journal at the time they published their article. (Cawte, Djagamara, and Barrett 1966). Drs. Singer and DeSole give a good review of subincision theories and correctly deduce that some weights as a determinant should be given to the grooved penis form of marsupials. As it stands, however, their article creates an unfortunate impression of their priority in this matter. Our article on subincision was not based solely on the classical literature and films, but on fresh fieldwork in central Australia. The present is a favorable time for this because with good fortune one may find men (such as our own collaborator Nan Djagamara) who are trifair English and have good rapport with Euro-