The previous chapter showed how, by collecting andscrutinising agricultural labourers’ testimonies,George Robertson made important contributions toJohn Phil[i]p Wood's Cramond parish survey. In fact,Wood and Robertson continued to collaborate andcorrespond – and, increasingly, to disagree –throughout the 1790s as they were both employed insending out the ‘Statistical Missionaries’. Thesewere questionnaires designed to prompt Statistical Accountcontributors to return their reports quickly and inthe desired format so that larger countycompilations could be published. Robertson wasresponsible for compiling distinct parish reportsinto the general county survey for Midlothian – alsoknown as Edinburghshire at the time – and he wroteoften to Wood to lament the challenges thisentailed. In 1796 Robertson complained of such a‘vast difficulty in extracting intelligence fromsome of the backward ministers, for example [theMidlothian parishes of] Colinton and Duddingston,that I am preparing to finish the account withoutthem. Indeed I would rather form a calculation fromaverages of the adjoining Parishes from theintelligence I have got other ways than plaguemyself more with them.’

This complaint is interesting in a number of ways.Firstly, there is the fact that Robertson was notcomplaining about the parish surveyors failing tocontribute at all; rather, he was angry that theyhad not contributed ‘intelligence’. Theircontributions were, in his opinion, either notvaluable or not accurate, or both. In fact, someministers simply provided data that contradictedRobertson's assumptions and hypotheses, as we willsee. Secondly, there is the language he used todescribe these unsatisfactory con-tributors: theywere a kind of ‘plague’ and, crucially, they were‘backward’. Theirs was apparently a failure to fullyappreciate the wider process of agricultural‘improvement’ of which Robertson was such a vocalspokesperson. Thirdly, it is notable that Robertsonthreatens to ‘finish the account’ of a county ofsome twenty parishes without including informationfrom at least 10 per cent of them, implying that, inhis view, the details from those few parishes wouldnot materially change the overall picture. This isrevealing of Robertson's view of local particularityand the function of place-specific data. Fourthly,and relatedly, it is striking that he plans to fillany gaps in ‘intelligence’ by generalising theaverages of the parish reports he was happywith.