Extract from: Meitzner, Laura S. and Martin L. Price. AMARANTH TO ZAI HOLES, Ideas for growing Food Under Difficult Conditions, Chapter 11 Human Health, ANTINUTRITIVE FACTORS AND PLANT TOXINS, Available: http://www.echotech.org/technical/az/aztext/azch11an.htm

RESEARCH. This brings us to an aspect of ECHO's ministry that is not normally visible to our network. Many undergraduate programs require research as part of the science major. ECHO encourages such students and their professors to undertake projects that would benefit small farmers in the third world. One of the projects we suggested in our "Research Opportunities" write-up was to look into this question of safety of velvet bean. Senior premedical major Sarah Kramer and her advisor Dr. Bob Kistler at Bethel College in Minnesota did just that, and came up with some very interesting information.

First, a computerized literature search turned up two journal article reports of people eating velvet beans. One study mentioned that they found a village in Ghana where some people ate velvet bean daily. Another study found that rural people in southern Nigeria use it as a soup thickener by first boiling to remove the hard seed coat, then grinding it.

Tom Post in Belize forwarded us a report like none other I have encountered from the book Poisonous Plants of the United States and Canada by J. M. Kingsbury. Using the velvet bean grown in Florida years ago "even boiled for an extended time, the beans were unpalatable and produced, an hour or more after ingestion, symptoms of nausea and discomfort. While cooking, the beans gave off a volatile substance which produced a smarting sensation in the eyes and a pronounced headache among those experimenting with them." This is so unlike recent reports where the tropical velvet bean is being used that there must be considerable differences in toxicity between varieties. ECHO distributes two varieties of velvet beans. One is the kind that has no itch-producing fuzz on the pods and produces seed only during short days. We call it our "tropical velvet bean." That is the one we normally send overseas unless specified differently. Seeds may be white, mottled or colored. The other is the less vigorous kind grown in the southeastern USA which we call the 90-day velvet bean and is possibly the kind mentioned in this report. However, Sarah's experiments with mice described below were with this 90-day type and she found no such problems.

Sarah's computer search turned up a rat feeding trial in Ghana using velvet beans. Results were reported in terms of grams of weight gain per gram of protein eaten (the protein efficiency ratio or PER). Rats fed raw beans lost weight (PER -3.03). The PER for rats fed autoclaved (i.e. pressure cooked) beans was 2.31, and for rats fed only the ideal diet it was 3.41. The lower value for beans does not necessarily mean there was still some toxicity. The protein of many legumes is not always digestible, or may be lower in one of the essential amino acids than the ideal control diet. The latter appears to be the case here because rats fed autoclaved beans to which the amino acid methionine (which is often in inadequate supply in legume seeds) was added had a PER of 3.59.

A study in the States showed that the likely benefit to the plant of such a high concentration of L-dopa is protection of the seed. "Mature seeds of velvet bean are conspicuously free from attack by small mammals and insects." Small amounts of L-dopa that they added to an insect diet produced toxic effects. Concentrations as high as found in velvet bean seeds inhibited feeding completely.

THE FEEDING EXPERIMENT. Sarah did a 27-day feeding trial with mice. She used the 90-day variety of velvet bean because we did not have enough of the tropical kind to do a feeding trial. Four mice were assigned to each of 9 experiments. The control mice were fed commercial mouse chow. When an experimental diet was used, every third day 4.0 grams of the control diet (mouse chow) was added to provide nutrients missing in the experimental diet. This amount was chosen because that was the average daily amount eaten by rats fed only the control diet. Mice fed the control diet gained 8 grams. Mice fed the control plus the amino acid methionine gained 7.5 grams, which statistically was not significantly different at the 1% confidence level.

[A note to those not familiar with statistics. In everyday English we use the word "significant" about the same as the phrase "a lot." "A Mercedes is significantly more expensive than a Volkswagon" means it costs "a lot" more. A scientist uses the word differently. If the weight of rats in two experiments is "significantly" different we mean that statistically speaking the probability is small that random chance could have accounted for the results.]

EFFECT OF "NUTRICAFE." One set of mice was fed the control diet except that velvet bean coffee was the only thing available to drink. Beans were roasted at 300øF on a cookie sheet for one hour. To make coffee, 40 g of roasted beans were boiled in 700 ml of water for one hour. Mice gained 10.2 grams, which was not significantly different from the control.

EFFECT OF RAW VELVET BEANS. Mice fed raw beans lost 5.6 grams. With added methionine they lost 6 grams. This could be caused by the L-dopa, but so many harmful things occur in raw bean seeds that some other cause cannot be ruled out. This was significantly different from the control.

EFFECT OF BOILING THE BEANS. Beans that had been soaked with one change of water were boiled 30 minutes (40 g in 700 ml water) then another 30 minutes in fresh water. They gained 4.8 grams. Mice fed boiled beans with added methionine gained 3.8 grams. These were not significantly different from the control.

For your consideration, Doug Welch in Malawi wrote, "Velvet beans are consumed here. They were displayed at the trade fair as one of the beans produced for consumption. There is a story of how villagers fled when attacked and left some partially cooked velvet beans. The hungry raiders ate them and all died. They have to cook the beans twice."

EFFECT OF ROASTING THE BEANS. Mice fed beans roasted at 300øF on a cookie sheet for one hour gained 1.5 grams. Those fed roasted beans plus methionine gained 3.0 grams. Both were significantly less than the control but not than the cooked beans.