When the weather turned hot and dry in August, many crop watchers said they thought that test weight, and hence yields, would be lowered by the dry weather. Reports coming in from many directions indicate that neither of these things happened in corn, though yields of soybeans harvested early have been disappointing and some extremely small soybean seed has been reported. Harvested soybean seed size has in some cases exceeded 5,000 seeds per pound, which is about half of normal size. This indicates that there were enough seeds for yields to be about twice what they ended up being; the crop simply ran out of water, with help in some cases from diseases that reduced root health or the ability of stems to transport water.
Once the number of seeds has been determined for either crop, yield is more or less completely a function of how large those seeds get. But final seed size often has little to do with test weight, and test weight really doesn't work as a way to explain the effects of stress on yield. While this may help some people make the connection between conditions and yield, it also implies to some people that low yields mean low test weight and that low test weights mean low yields. That is not the case, at least in a specific sense.
When early frost or something else quickly cuts off grain filling, corn usually has both low yields and low test weight. That's because starch formation stops quickly, so density of the endosperm ends up being low. Low endosperm (kernel) density usually means low test weight, sort of like Ping Pong balls compared to golf balls. But test weight isn't simply a measure of kernel density; it also depends on how well kernels fit together (size and shape), how slippery their seedcoats are, and other physical characteristics.
Once kernel density reaches a certain point, other characteristics of the seed determine test weight. In fact, seeds that fill for a little extra time and so get heavier could even lose test weight if their shape changes in the process. Likewise, moisture changing from 28 to 20 usually increases test weight because the kernels physically shrink and get more dense, while changing moisture from 20 to 15 can actually decrease test weight, because kernels can lose moisture (weight) without changing their size or shape.
Corn yield is number of kernels per acre times weight per kernel at standard moisture divided by 56 pounds per bushel to give bushels per acre. Test weight is a measure of how many pounds of corn will fit into a certain volume and is only partly related to weight per kernel. Popcorn tends to have lower yields and high test weights. Sweet corn fills poorly and so has low yields and very low test weights. Field corn can go either way. But test weights above 60 are very unusual in field corn, no matter whether it yields 125 or 250 bushels per acre.
This year, kernel number was usually set before stress began, and so whether kernels got larger or stopped filling early had a big effect on yield (weight per kernel) but not always much effect on test weight. I think we could say for certain that yield and test weight will not change by equal percentages and that there is no formula to predict how they will be correlated. Test weight is still used as a sort of "quality" measure by breeders, though, because it does have some correlation with kernel density and it is easy to measure.
Standard test weights for different crops go back to the time when people measured yields by volume, not by weight. A bushel as it was first used hundreds of years ago was a volume measure (1.24 cubic feet), used because people either couldn't weigh things accurately or they couldn't be trusted to do so. A volume measure was a good answer, because the customer could easily see whether the measure was off just by looking at its dimensions. Length was much easier to measure than weight.
Today, we sell corn in 56-pound "chunks" and soybean in 60-pound "chunks" that we still call bushels; but except for wanting to know how many "bushels" will fit into a bin, we don't use it as a volume measure as they did in the old days. I think that is where the "test weight = yield" thinking comes from, though.
Soybean test weight doesn't seem to vary that much, and it doesn't get much attention. Soybean seeds tend to be more regular in shape than corn seeds, and spherical shapes like soybean seed tend to fit together more uniformly regardless of size differences. A shortened seed-filling period may also result in smaller seed but not much change in density of individual seeds.
Anything--dry weather, clouds, cool temperatures, insects, diseases--that reduces the rate or seasonal duration of photosynthesis will reduce the rate of seed filling in any crop and so will reduce yield unless the season is lengthened to compensate. Any such yield reduction might lower test weight or it might not, but the statement that yields will be lower "because of lower test weights" isn't the way it works. In terms of cause and effect, it is more accurate to say that test weights might be lower as a result of lower yields, given that yield changes this late in the season are almost entirely a result of changes in kernel weight.--Emerson Nafziger