The machine measures and records every stem it takes. The file says 212. Two men walked the block with tally counters and found 211. The mill weighed the load and the mill agrees with the machine.
I have operated a harvester for nineteen years. A harvester is the machine with the boom and the head on the end of it that grabs a standing tree, fells it, strips the branches off and cuts it into lengths, all in about fifteen seconds.
Nobody believes the part that matters, so I will do that first.
The head is full of sensors. It measures the diameter of the stem with a pair of knives that close onto it, and it measures length with a toothed wheel that runs along the trunk as the stem is fed through. It has to, because that is how it decides where to cut: a sawmill wants a log of a given length and minimum top diameter, and the machine works out the best combination of logs from the tree it is holding, while it is holding it.
All of that is written down. Every stem gets a record: when it was cut, its diameter at the butt, its length, the logs made from it, the volume. The file is a standard format the whole industry uses and the forest owner gets it, because the owner is paid on the machine's own measurement.
So I do not have an opinion about how many trees I cut. I have a file.
Compartment 9 is a twenty hectare block of Sitka on a slope, thinned three times and clear felled in rotation, and I have worked in it on and off since 2011.
The first time was a routine thinning and the routine is this: I cut, the file goes to the office, and two of the ground crew walk the racks afterwards with tally counters and count stumps, because that is the check. The machine can double count a stem if the head slips, and a stump is a stump.
The file said two hundred and twelve stems.
The count said two hundred and eleven.
I assumed the counters had missed one, which is what anybody would assume, because walking a thinned rack counting stumps in brash up to your knee is miserable work and one in two hundred is a good error rate.
They went back and did it again. Two hundred and eleven.
So I went and did it myself, on a Saturday, in my own time, with a can of marker paint, and I sprayed every stump I counted so I could not count one twice.
Two hundred and eleven.
Here is what makes it not a counting problem.
The load went to the mill. A mill weighs everything that arrives, because that is the other way timber is paid for, and the weight of a lorry load of fresh Sitka is predictable to within a few per cent of its volume.
The mill's weight matched the machine's volume.
Two hundred and twelve stems' worth of timber left that block and was weighed by somebody with no interest in the argument, and two hundred and eleven stumps are in the ground, and they have been counted four times by three people, one of whom painted them.
What the file says about the extra one.
I can tell you exactly which record it is, because after the second time it happened I started looking.
It is never the first stem of the day and never the last. It sits in the middle of a run, with a timestamp eleven to forty seconds after the stem before it, which is normal working rhythm. Butt diameter in the thirties of centimetres, which is normal for the stand. Two logs made from it, both to spec. Nothing flagged.
It is, in every way the file can express, an ordinary tree that I cut.
I do not remember cutting it. I want to be careful there, because I cut four hundred trees on a good day and I remember almost none of them individually, so not remembering is worth nothing as evidence and I am not offering it as any.
It has happened six times.
Every thinning and every fell in Compartment 9 since 2011. Six operations, six files, six stump counts.
212 against 211. 389 against 388. 514 against 513. 166 against 165. 431 against 430. 908 against 907.
One. Every time. Never two, never none.
And it does not happen anywhere else. I have worked thirty or forty blocks on this estate and the counts agree or they disagree by a handful in both directions, the way an error does. Compartment 9 is short by exactly one stump and it has been for fifteen years.
Things I have checked, because I have had a long time in a quiet cab to think about what to check.
The head does not double count. A double count leaves two records with nearly identical diameters a second or two apart, and I have looked at all six files for that pattern and it is not there.
The boundary. An obvious answer is that I am reaching over the line and taking a tree out of Compartment 8, so the stump is in the next block and nobody counts it. I measured. The boom is ten metres and I have walked the boundary with the GPS on the machine and there is a twelve metre unstocked strip along it where a ride used to be. There is nothing within reach to steal.
Trees that were already down. A windblown stem can be processed without making a stump. There is none in the files: every record has a butt diameter measured by the knives, which means the machine held it standing.
The counters. Three different people, four counts, paint on every stump.
The thing I do now, which I have told nobody at work.
On the last operation in there I watched the counter on the display instead of the trees. It runs up as you work and I have never once looked at it in nineteen years, because what would you look at it for.
At two hundred and twelve it read two hundred and twelve. I was holding a stem, the head was feeding, and everything was where it should be.
The stem after it made the count two hundred and fourteen.
I stopped, and I put the boom down, and I sat there for a while with the engine running, and then I did what you do, which is carry on, because there were four hours of daylight left and a forwarder behind me waiting for timber.
When the file came off the machine that evening it was clean. Every number in sequence, nothing missing, nothing doubled, and the total at the bottom was nine hundred and eight.
They counted nine hundred and seven.
I am fifty-two and I have six files and six counts and a can of paint, and the only thing I have ever worked out is what it is not.
The estate has Compartment 9 down for clear fell in 2029, which is three years, and after that it is replanted and the stumps go under a mulcher and there is nothing left to count.
I have put in for a different block that year. I have not said why and nobody has asked, because a man asking for easier ground at fifty-two is the most ordinary request in this industry.