It’s a century, this weekend, since New Zealand forces attacked Chunuk Bair as part of a failed effort to end the Gallipoli campaign. Curiously, we neither know exactly how many New Zealanders fought in that eight-month campaign – or how many became casualties.
That question has been exercising some of the key figures in New Zealand’s military-historical community in the past few weeks, all behind the scenes. Historiographical discussion is always useful because, if constructive, it can often tease out understanding in unexpected ways. I won’t detail all the emails, but the latest debate was triggered by an exhibition in New Zealand’s national museum, Te Papa Tongarewa, The Scale Of Our War – built by Weta Workshop – which includes an infographic giving a figure of 93 percent for the New Zealand casualty rate during the eight-month Gallipoli campaign.
That figure derives from the usual numbers given for our participation – 8556 – against the casualty figure (wounds, sickness, death) of 7447. It works out at 87 percent on those numbers, from an iconoclastic foreword by General Sir Ian Hamilton to Fred Waite’s Gallipoli history of 1919. Other figures, added to it, give the 93 percent value.
The problem is that while First World War casualty rates were colossal, they usually weren’t that high. New Zealand’s average, across the whole span of the war, including the Western Front 1916-18, was 58 percent killed and wounded. And the tactical problems that provoked that rate – the supremacy of machine-gun-and-wire defence over infantry offence on a scale of battlefield defined by the range of those guns – were similar in all theatres where trench warfare occurred.
Australia’s equivalent rate in Gallipoli was 60 percent, the other nations less. So how did New Zealand, alone, end up with a higher rate at Gallipoli, where the tactics and technologies were the same, and where nobody else in that theatre suffered so heavily? Statistically, 87-93 percent is an outlier.
The historian behind this exhibition is a friend of mine – Christopher Pugsley. In my opinion he’s New Zealand’s top military historian – he’s defined and re-defined the field so completely that the rest of us follow in his wake. I’m aware of his reasoning, because he sent me the paper he wrote on it.
My own take is that while Hamilton’s number has been the officially accepted figure for decades, circumstantial evidence points to a higher participation rate. Efforts have been made to identify it. However, nailing that figure – and proving it – is fraught with complexity, because accurate data doesn’t really exist.
I can’t do more here than briefly summarise some of the issues – there’s more detail than I can discuss in a blog post, and what I’m focussing on here are the terms of debate.
First port of call, as far as I am concerned, are Hamilton’s numbers. Historian David Green argues that they likely derived from figures originally collected for Waite’s book, adding the Main Body of the Expeditionary Force to others raised in Egypt for a total of 8556. Waite’s book contains an appendix showing those numbers, and states that these don’t include later reinforcements sent to Gallipoli. It’s a compelling argument, given that the numbers correlate. The missing empirical link is that Hamilton, by his own admission, hadn’t seen Waite’s book, and nobody’s found specific documents showing what he did see in order to get that number.
The other part of the 87-93% rate flows from the casualty figure. This, too, is uncertain, because some soldiers were withdrawn wounded, returned to Gallipoli and were then wounded again or killed – meaning they were listed twice (or more), inflating the casualty figures. Even at the time, officials doubted the accuracy of casualty data reaching the New Zealand Records office in Cairo, staffed with men from the 3rd Echelon. MP Heaton Rhodes discovered the point during an official visit to Egypt in November 1915, finding records were particularly bad for the first months of the operation. ‘Mistakes …must be attributed to the difficulty of obtaining accurate information from the units, the members of which were then fighting for their own lives. I am sure they did everything possible to report casualties correctly, but, mistakes were unavoidable owing to stress of action and the difficulties of communication…’
Can the potentially understated participation rates and overstated casualties be proven empirically? Therein is the problem.
We know that roughly 20,000 New Zealanders had been despatched to Egypt by late 1915. Their records are available. However, there are reasons why examining them won’t give definitive answers. Where a record shows despatch to Gallipoli it’s clear enough. But some records show despatch to Mudros (Moudros) – the usual assembly point and forward base for the campaign – but not Gallipoli. And yet we know that soldiers despatched to Mudros often ended up on Gallipoli without that being noted.
I have a family example: the record of my great uncle, Corporal Ben Draper, DCM, 11/1255 (1894-1982), shows he was posted to Mudros on 25 October 1915. Nothing is shown of a transfer to Gallipoli. However, I know my great uncle went to Gallipoli – not just because he said so – but because his record shows him reaching Alexandria from Anzac on 25 December 1915, meaning he was part of the evacuation at the end of the campaign. So either Uncle Ben teleported to Gallipoli from Mudros, or army book-keepers regarded them as effectively the same destination.
I suspect what we’re seeing is evidence of a subculture. Army officials didn’t keep records for history. They kept them for their own purposes. Mudros was one of the assembly areas, and it was almost (but not quite) axiomatic that a soldier sent there would end up in Gallipoli, so the final journey was often not listed. Soldiers wounded or taken ill at Anzac were often evacuated to Mudros – and that was shown in their casualty record – but if they weren’t, there wouldn’t be anything official to show they’d been on Gallipoli other than, maybe, the record showing their return to Egypt, typified by my great uncle’s papers.
But the application by period officials seems to vary, creating a level of doubt now.
What’s concerned me about the argument so far is that terms of debate have been framed within the parameters of ‘history’ as an analytical discipline within the humanities – a ‘soft’ study which lacks a methodology to properly understand and transcend numerical data limits. However, history is only one of the things I do, and I could see an obvious way of cheaply answering the problem via mathematical technique.
So the other week I caught up with New Zealand’s chief government historian, Neill Atkinson, who I’ve known professionally for some years, to discuss the issue. As I explained to Neill – why not try:
- A statistical analysis of a random sample of the 20,000 soldiers’ records, showing whereabouts between April and December 1915 and any casualty figures (if listed), correlated with location, taking special note of destination (Mudros or Anzac) and correlating that with any evidence that the soldier was, in fact, on Gallipoli even if they were sent to Mudros – as my great uncle’s record clearly shows. To be valid, this selection must be random, not taken from, e.g. a battalion for which complete records are available. Bayesian regression analysis applies.
- Analysis of numbers sent by New Zealand to Egypt or the Mediterranean theatre in 1914-15, giving baseline ranges for the potential total available for service in Gallipoli.
- A statistical analysis of New Zealand’s overall participation rate vs casualties for the whole of the First World War, compared against selected other nations (noting that most nations have variable records in this regard).
- Ethnographic search of randomly selected soldiers’ diaries and records from the whole of the NZEF, seeking comparative reference to excessive loss rates compared to other nations; or to the ‘safer’ environment of the Western Front compared to Gallipoli. A search should also be made for descriptions of troop levels left at base in Egypt, Zeitoun camp. Negative results are still positive for the purpose of this exercise.
This will still produce a range of figures, with defined uncertainties, but it will do so in a structured way. We need to accept that the answer may be ‘nobody knows’. In that case, it’s fair to state the range with the qualifier that the data is incomplete, to state the values of the uncertainties, and the logic around them. The answer, in short, is likely to be an essay rather than a number, with critical source evaluation and statistical analysis. But, philosophically, this is also a valid answer.
I will not, of course, be doing this work myself. I’ve long since ceased to seek or value status with the New Zealand academy. Nor do I seek it with New Zealand’s military-historical community, some of whom are good friends, but others of whom have behaved towards me in a way that has materially failed to earn my trust and respect. Besides, I’m busy with other things – not just my back-list re-release programme but also an upcoming hard sci-fi novella. And a book about the New Zealand soldier in the First World War 1914-18 – a novel. With zombies.
Copyright © Matthew Wright 2015