It is a one-sided presentation of the facts, with little reference to the bestiality and horror of the Mau Mau oath taking ceremonies, or the force used on those unwilling to take the oaths. Nor does she bring out that Mau Mau was opposed by all Kikuyu Christians, many of whom were killed - as in the Lari massacre - for their opposition.

Much of the so-called evidence comes from interviews with ex-detainees with little opportunity to check their veracity. This leads to some major errors. A good example is at p313 where, on the basis of an interview with one Nderi, she states that Leslie Whitehouse, DC Turkana, "handed out torture". It appears from the bibliography that she has not read the biography of Whitehouse (Jomo 's Jailor: Grand Warrior of Kenya by Elizabeth Watkins) from which she could have seen that Whitehouse was a fair-minded, gentle, and conscientious administrator who would never have tortured anyone. He became a good friend of Jomo Kenyatta, became a Kenya citizen, was asked to stand as a KANU candidate, and was awarded the decoration of Grand Warrior of Kenya by an independent Kenya. It is fortunate for her that Whitehouse is now dead, as otherwise I suspect he could have won a very substantial sum in a libel action.

Some errors were made at Hola - and fully aired in the British Parliament by the parliamentary opposition. But they do not justify the all-out condemnation of those who worked in the detention camps, nor does Elkins suggest how else men could have been freed from these horrific oaths. There were many who worked patiently to help detainees get free of their oaths and live normal lives. Two of these were James Breckenridge and his wife who describe their efforts in Breckenridge's book 40 years in Kenya (another book that does not appear in Elkins' bibliography).

"Elkins is a prominent political activist who for some time has been campaigning for monetary compensation for the alleged victims of British efforts at defeating Mau Mau. Yet, as the author of this work, she has attempted to drape herself uniquely in an academic mantle, namely her assistant professorship in Harvard's history department. There is obvious subterfuge here: In failing to inform readers of her primary role as a political activist, she has attempted to camouflage the bias this clearly imparts to her historical narrative." ...

"Thus, through inappropriate analogies and inflammatory rhetoric, Elkins attempts not so much to present truth supported by incontrovertible evidence, but rather to solicit broad public support for her crusade on behalf of Mau Mau adherents and sympathizers who were detained. In effect, her book is less a serious scholarly narrative and more a political brief crafted in popular language in the interests of a group of people whom she considers victims of a past wrong and worthy of reparations." ...

"A major defect in Imperial Reckoning is the bias and subjectivity with which Elkins has researched the available evidence and drawn conclusions from it." ...

"In the end. Imperial Reckoning rests heavily on fragile testimony, faulty data, and unconvincing intuition. It informs readers more about Elkins and her crusade than about the complexities of Mau Mau and the human tragedies associated with it."

Drawing on my experience of Kenya in and after the Emergency and my later career as a market researcher for 33 years I would summarise his conclusions and add my comments as follows:

1. Caroline Elkins' assertions that up to 300,000 Kikuyu (here and later this definition subsumes Meru, Embu, Tharaka and Mbere) died or disappeared as a result of the Mau Mau Emergency (including those not born because of a lower than normal Kikuyu birthrate) is disproved. There may be some truth in her claim "that the official figure of some 11,000 Mau Mau killed is implausible". (Censuses, of course, even if perfectly designed and carried out, with 100% coverage, co-operation, and understanding from the population concerned, are not intended to ascertain the exact size of groups as small as one seven hundred and fiftieth of the total population. More realistically they should be able to establish whether a defined category of person amounts to a thirtieth, or to only a fortieth or fiftieth, ie two to two and a half percent. Greater precision is seldom necessary for economic and other planning purposes, but there may be political imperatives for it).

2. Elkins has ignored changes in the tribal classifications used for the Kenya censuses. More importantly she has stressed apparent differences in population growth between the Kikuyu and three other major groups, the Luo, Luhya and Kamba, implying that this is accounted for by mass killings or is the consequence of other aspects of colonial, ie British or Kenya, government policy. Blacker shows that Kikuyu mortality, for both adults and children, was appreciably lower than among the Kamba, Luhya and Luo: this pattern has been repeated in all subsequent censuses and health studies. While conceding that there may have been malnutrition in Kikuyu areas, aggravated by villagisation and the spread of infectious diseases (it would be interesting to obtain a Red Cross or medical view on this: Blacker surprisingly quotes Elkins as his source), the author points out the relevance of environmental (climate, altitude, malaria) and social (education, age at marriage and at birth of first child) factors as reasons for tribal differences and concludes that "... the excess mortality of the Kikuyu attributable to the Emergency cannot possibly be measured by comparisons with other tribes."

3. So what was this "excess mortality"? Blacker has calculated this from a comparison between rather incomplete 1948 data obtained from women on the number of children borne, distinguishing between the living and the dead, and the fuller answers collected in the 1969 census, reinforced by a modelling procedure to infer the number of deaths at all other ages. He concedes "the fragile nature of the data and assumptions" and the "large margins of error" involved, hence his range of 30,000 to 60,000 "excess deaths" and his statement that "a round figure of 50,000 is as good a guess as any." Of these "... rather more than half, say 26,000, will have been of children under 10 ...". The estimates provide "an order of magnitude" very different from Elkins' claims. Within the total, about 15,000 violent deaths already known about and admitted are included.

4. A few further observations:

i) The casualty figures in the Corfield Report may be official but they are necessarily incomplete, only reporting up to the end of 1956 and appearing to omit the KEM (Home) Guard losses and possibly the Kenya Regiment's African trackers. There was (and has been) no attempt to estimate Mau Mau killings of their own men within the gangs (or "armies").

ii) Some Mau Mau continued to live in (or return to) the forests, particularly in Mem, after military operations ceased (1956). Nearly 2,000 appeared at the time of independence (1963). Would they have participated in the 1962 census? Police operations continued for the rest of the fifties and in the early sixties, including after independence.

iii) 59,000 "Non-Kenya Africans" were identified in the 1962 census. What we don't know is the number of Kenya Africans (particularly Kikuyu) living outside Kenya in that year: I would expect the total to exceed what was recorded in the 1948 Tanganyika and Uganda censuses: these figures should still be available.

iv) Overall under-enumeration in the 1948 census probably exceeded that for 1962 but the reverse may be true for the Kikuyu, but might there not have been some incentive to exaggerate numbers to improve a tribe's or district's representation in the forthcoming 1963 elections, counter-balancing Blacker's suggestion that Kikuyu resentment and non-cooperation may have contributed to underenumeration in 1962?

v) Dr Blacker has confirmed that the published article contained an error on p 223, Table 10, where the enumerated figures for "both sexes" and "males" should be as shown in Table 8, being 1,554,925 and 762,536 respectively. The "females" figure is correct, and his conclusions are not affected.