2008 harvest report

Well, over 2 tonnes of fruit were harvested over about five weeks from about 10th October. This includes on- (75%) and off-farm (25%) sources. On-farm, harvesting of ‘3’ trees was independent of ‘1’ and ‘2’ trees (see earlier post on range scores). All nuts from fruiting trees were counted, though 1 and 2 trees were counted at time of pick up (three different sweeps of the fields), and the fruit pooled by field during each sweep. Fruit from each 3 tree was kept separate in plastic lugs in the field, and totals counted when there were no further nuts to fall.

All pooled fruit was processed first – weighed, hulled, then washed and dried. Fruit from 3 trees was weighed, hulled, washed, then sampled and dried. Three-nut samples of 3 fruit were retained for drying, then subsequent cracking and separating. This analysis will be reported under the 2008 Nut Report.

Within the global number reported above, there was much hidden. The two main experimental populations (Fields 1 & 3) are reported below (you can click on the image to open it in a larger format).

Firstly, the nut range scoring system was successful in separating trees by eventual yield (at harvest). In each of the two fields, there are in excess of 500 trees. In each field, more than half of the population did not fruit. The number of trees fruiting (total n) was 187 (F1) and 254 (F3). The average number of nuts per tree was 20 (R1), 94 (R2) and 299 (R3). The biggest contrast between fields was for R3, where F1 gave 448 nuts per tree and F3 gave 151 nuts per tree. Correction of these totals by cross-sectional area (CSA) at DBH (diameter at breast height; multiplied by 100 in the above graph to be able to use the same vertical axis) almost completely reduced this variability to 5.09 nuts per sq cm of CSA (F1) compared to 5.53 (F3). I have termed this value Nut Yield (NY), which you can see was, on average, >1 for R1 and between 2-3 for R2 (remember the 100 factor in the graph above). Groups of bars for same range score number are from fields 1 and 3, reading from the left.

So what? (you might say). Well, it confirmed that it is not just tree size (measured here as DBH) which accounts for NY. There are other factors which make some trees more productive than others. It also suggests that you could forget about harvesting R1 trees and still collect just about the same total weight of fruit, or, put another way, R1 trees contribute very little, individually, to the total harvest, though they make some of this up by sheer numbers. As we are embarked on an inter-generational venture, as long as NY is the main contributor to total kernel yield (KY) per tree, it is only from R3 trees from which we should wish to select. We would want to analyze this characteristic only within the population of R3 trees. It also offers the means (i.e. a sufficient number of very uniform nuts) to set a cracking machine to crack optimally for a given set of nut characteristics, which is some of the rationale underlying US selection of black walnut ‘cultivars’ based on kernel percentage (K%; sometimes measured in efficiency terms as ‘crack out’), though high K% is the other aspect.

In my top-yielding F1 tree, nut number was >1700, with a NY of 7.88. However the top NY (16.15) was from the top F3 tree which gave only 343 nuts. Correction for size is thus immensely important to understanding productivity within a relatively uniform environment, and were both trees to show the same K%, the first would be of greater economic importance to me at this point in time. The former had a 2007 K% of 26% in a nut weight of 15g, which suggests a total KY of 6.4kg from this tree this year. Not insignificant.

Why do I think so many trees are not producing? Probably because they are too early for my site. I shall test this next spring, comparing a random sample with the ‘3’ trees.

Finally, you probably think I’m crazy to do so much counting. Learn to do it, for it will tell you a lot. Use range scoring first, though, so you only count nuts from the ‘3’ trees.