Cold Room Caliper Measures Skulls

The craniometer arrived from Cambridge on the ninth of March, 1893, and I unwrapped it in the cold room at University College where the skulls from the Naples collection sit in their numbered trays. A brass sliding caliper, vernier to a tenth of a millimetre. For three years I have measured these crania with a wooden rule and a pair of dividers, recording each in the ledger to the nearest half-millimetre. The old numbers are in the book. The new instrument says they are wrong.

Not wrong by much. Wrong by enough. A skull I set at 183.0 in 1891 reads 182.4 on the vernier. Multiply that error across three hundred specimens and the correlation I published between cephalic index and stature shifts in the third decimal - trivial, one says, except that the third decimal is where I have been arguing my case. The instrument does not care about my argument. It measures.

Here is the assumption I trust least, and I write it down because writing it down is the only way to see it clearly: I have assumed the error is random. That it scatters around the true value and cancels in the aggregate. But the old rule was read by the same hand, my hand, and a hand has a bias. If I held the caliper at a slight angle, if I favoured the occipital ridge, the error is not random. It is systematic. And a systematic error does not cancel. It accumulates. It tilts the whole distribution.

I cannot test this without remeasuring everything. So I will remeasure everything. Starting Monday, the full Naples series, both instruments side by side, and I will record the difference for each specimen, not the value. The difference is the datum now.