The internet has been a bit a-flutter recently on the subject of declining nutrients, particularly in fruit and veg. We nibbled as much a week or so ago, and I asked whether the data had been published. 1 Indeed it had, as Amanda Rose graciously pointed out. As for reasons, Amanda said:
An editor of Organic Gardening magazine suggested that use of chemical fertilizers and subsequent decline in soil minerals was the cause. A peer reviewed study of the data provides another explanation: commercial cultivation of seed has traded nutrient density for yield, pest resistance, and transportability, factors critical to the commercial success of any crop.
Off, then, to the peer reviewed study. The best I can say is: it’s complex. Don Davis, of the University of Texas, and his colleagues used existing USDA data on the nutrient content of 43 garden crops, mostly vegetables, to compare quantities from the 1950 edition of the USDA’s “Composition of Foods” with those in the 1999 edition. Many pitfalls await the unwary in such an exercise, not least of which is the lack of basic information about some of the sample sizes and distribution of the results. Davis and his colleagues seem to have thought of them all, and worked round them as far as possible in various ways.2 In essence, Davis et. al compare median values in 1950 with medians in 1999; less than 1 represents a decrease, more than 1 an increase. And the bottom line:
As a group, the 43 foods show apparent, statistically reliable declines (R < 1) for 6 nutrients (protein, Ca, P, Fe, riboflavin and ascorbic acid), but no statistically reliable changes for 7 other nutrients. Declines in the medians range from 6% for protein to 38% for riboflavin. When evaluated for individual foods and nutrients, R-values are usually not distinguishable from 1 with current data. Depending on whether we use low or high estimates of the 1950 SEs, respectively 33% or 20% of the apparent R-values differ reliably from 1. Significantly, about 28% of these R-values exceed 1.
That needs unpacking. First, there’s the question of “apparent” declines. Davis points out that while his statistical approaches eliminate random uncertainties, it is always possible to postulate a systematic error affecting any particular nutrient in either direction. Iron, for example, is tens of thousands of times more abundant in soil than in crops; merely changing the way samples are washed, to remove more soil in the 1999 samples, would result in a large decrease in iron.
That is why I call these R values “apparent”.
But there’s a kicker:
However, it would seem scarcely credible to attribute all the statistically significant median R < 1 to multiple systematic errors, each one operating in only one direction.
So the overall decline, considering all the crops and all the nutrients, is real enough, even though for each crop and each nutrient most of the differences (between 67% and 80%) are not statistically significant.
Some nutrients did increase, most notably vitamin A and riboflavin. Which makes it plenty delicious that in all the fluttering about declining nutrients, Scientific American (no less) chose to major on carrots,3 and to blame it all on “soil depletion”.
You mean it isn’t soil depletion?
Davis comes down firmly against the “organic” idea that “chemical fertilizers and subsequent decline in soil minerals” is the problem. The fact that some nutrients definitely increased, coupled with the observation that protein (mostly nitrogen, N), phosphorus (P) and ash (mostly potassium, K) each declined from 1950 to 1999 – despite the fact that “chemical fertilizers” are largely N, P and K – puts paid to that idea.
Instead, Davis et al. focus on two different “dilution” effects. Environmental dilution is an idea that has been around a while. In essence, plants in well-fertilized, well-watered soil grow larger but take up the same total nutrients. So the nutrients are spread through a larger crop, giving a lower concentration per gram of dry matter.
Davis also identified another effect that he called genetic dilution. The bulk of the dry weight yield of most fruits, vegetables and grains is carbohydrate. When breeders select for high yield, they are effectively selecting for high carbohydrates …
… with no assurance that dozens of other nutrients and thousands of phytochemicals will all increase in proportion to yield. Thus, genetic dilution effects seem unsurprising.
The best evidence for genetic dilution comes from contemporary studies in which a time-series of varieties is grown side by side, eliminating all the problems of historical data, agronomic practices, etc etc. Davis cites a few such studies, and more are being published. In general, they do show declines in nutrient density, which can be ascribed to breeder selection.
So what about the question Luigi posed a while back? Is modern plant breeding bad for your health? Possibly, a little, but not in the way most people think. The big problem isn’t that nutrients in fruit and vegetables have decreased; despite the decline, they remain among the most nutrient dense foods you can eat. So eat them. Choosing one growing regime rather than another isn’t going to make any difference. The big problem is that we’re still eating too little fruit and veg and too much refined staples. Last word to Don Davis:
Our findings give one more reason to eat more vegetables and fruits, because for nearly all nutrients they remain our most nutrient-dense foods. Our findings also give one more reason to eat fewer refined foods (added sugars, added fats and oils, and white flour and rice), because their refining causes much deeper and broader nutrient losses than the declines we find for garden crops.
Technology should allow us to increase selected nutrient concentrations.4 But will we learn 20 or 40 years later that there were new, unintended side effects? Another question looms large: Is it wise, in the era of technology, to keep crop size (or even the concentrations of a few, selected nutrients) as our primary measure of farming success?
- Fully intending this post to be a snit-fest; there’s a bit of snit, I confess, but the subject is too important to lose to bitterness and acrimony. And the main researcher concerned can fully take care of himself. [↩]
- The statistics are not for the faint-hearted except that, bless them, they preferred non-parametric over parametric tests. That I can relate to. [↩]
- Vitamin a: 1950=12417, 1999=28129. [↩]
- As, indeed, it has: witness HarvestPlus’ spiffy iron-man pearl millet. [↩]