Table 6.0
Carbohydrate (Sugar) Profile Percentages:
Milks vs. Fruit Blends

Units: % of total available carbohydrates in profile,
for each food item.

Table 7.0
Fatty Acid Profile Percentages:
Milks vs. Fruit Blends

Units: % of total fatty acids in profile,
for each food item.

What we are presented with in the form of Tables 5.0, 6.0, and 7.0 above is a list of numbers for human milk, goat milk, and two fruit blends. Which of the other lists is "closest" to the one for human milk? One approach to this question is to use a simple statistical technique: ordinary least squares, or linear regression, to compare the data.

Recall that we are interested in the fruitarian claim that "fruit is like mother's milk." An easy way to test this, in the context of the amino acid and fatty acid profiles, is to fit a series of simple linear models (straight lines) where the dependent variable is the amino/fatty acid profile for human milk, and the independent variable is, in turn, the equivalent profile for goat milk, and each of the two fruit blends. (Note: The entire nutritional profile cannot be compared using this technique, as the entire profile consists of different types of data in different units--a dissimilar data set.)

The results of these model fits are given below. The merit of the results, or the model fit, can be evaluated via the F and R-square statistics below. The F-statistic compares the sums of squares ("variability" in crude terms) for the regression model (fit) versus the sums of squares for the error term in the model. It provides one assessment of the goodness of fit of the models. The degrees of freedom for the F-statistic reflects the number of independent pieces of information involving the dependent variable used in its computation: The numerator is for the sum of squares of the regression equation, and the denominator for the sum of squares of the error term in the model. The "Prob > F" column gives the probability of obtaining an F statistic larger than the one listed for the regression, by pure chance. Typically, one wants this number to be 0.05 or less--this is the de facto standard.

The R-square statistic measures the amount of the variability in the data that is explained by the linear model. It varies from 0 to 1. There are no hard rules regarding which values of R-square are acceptable, and which are not, though there is general agreement that values close to 1 are "good," and values that are "low" are "bad." Here, I will use a conservative approach, and consider values of R-square of 0.9 or higher to be acceptable, and a good fit, while values much below 0.9 are considered to be a poor model fit and/or unacceptable (i.e., "not a good fit").

To be precise, we can consider the model fit as good/acceptable if BOTH conditions are met:

1. An F-statistic for which Prob > F is 0.05 or less,
   --AND--
   
2. an R-square statistic which is 0.9 or higher.

Table 8.0
Amino Acid Profiles:
Human Milk Fitted as a Function of Profile(X)

Table 9.0
Fatty Acid Profiles:
Human Milk Fitted as a Function of Profile(X)

Side note: The question of interpreting the significance of regressions is discussed in N.R. Draper and H. Smith (1981) Applied Regression Analysis, Wiley: New York. Draper and Smith suggest that the F-statistic be at least 4-5 times the minimum value for acceptance at the P-level in use (P=0.05 level); such a criteria would lead us to reject some of the above regressions. However, use of the R-square statistic leads us to a similar result.

The paragraph below is from the article,

"Uncommon Uses of Avocado," by Yasseen Mohamed-Yasseen, Ph.D., University of Florida TREC [Tropical Research and Education Center, Homestead, Florida], appearing in: Tropical Fruit News, 28(1), Jan. 1994, p. 14; published by Rare Fruit Council International, Inc.: Miami, Florida.

Avocados are the richest known natural source of the seven-carbon sugar D-manno-heptulose [Simon and Kracier 1996; Otaga et al. 1972]. This sugar possesses the physiological ability to cause inhibition of insulin secretion in humans, thereby producing "instant diabetes" [Simon and Kracier 1996]. It is improbable that blood sugar levels in a normal person would be affected by average consumption; however, diabetics should consume avocado cautiously [Bergh 1992].

Comments on Mohamed-Yasseen:

The suggestion by Mohamed-Yasseen that diabetics should consume avocados cautiously because they may suppress insulin production raises questions for fruitarians as well. First, given that most fruitarian diets are high in sugar from sweet fruit, one should be cautious about introducing a food that inhibits insulin production.

Second, those whose diet includes large amounts of sweet foods, particularly fruit, often experience, on an ongoing (chronic) basis, the symptoms of diabetes: excess urination, thirst, sugar highs/lows (mood and energy swings), intermittent blurred vision, fatigue, pains in large joints, etc. This suggests (but is not hard proof) that excess sweet fruit consumption may induce a diabetes-like disorder that operates as long as the sweet fruit consumption continues.

Third, avocados are the prime fat source (and calorie source) for many fruitarians. Despite the "party line" that they should be eaten in small amounts, many rawists eat the equivalent of 1.5-2 or more large avocados per day. (Despite the contrary claims of a few extremists, fat is an essential nutrient. The body knows this even if the ego does not, hence the "cravings"--actual hunger--for fat in the form of avocados.)

Further, Otaga et al. state (p. 115): "It has been well demonstrated that small doses [of mannoheptulose] which do not produce hyperglycemia still produce measurable changes in the secretion of insulin." In the context of this paper, this suggests that those who choose to live on a diet of sweet fruit and avocado (like the blend analyzed here), should be cautious about how/when the avocado is consumed. It might be best to eat avocado separately from sweet/starchy foods. (Interesting--this might be one of the few cases where food combining has real scientific merit.)

From the above analysis, we see that:

• Human and goat milk are both fatty foods, while more than 50% of the calories in the fruit blends are from sugar.

• The vitamin and mineral profiles for human and goat milk are generally comparable, and different from the profiles for the fruit blends.

• Human and goat milk both contain as primary sugar lactose (which is slowly absorbed), while sweet fruits contain fructose and rapidly assimilated sugars (glucose, sucrose) as primary carbohydrates.

• The amino acid profile for goat milk provides a good statistical "explanation" (degree of relative fit) for the amino acid profile for human milk. The fruit blends do not explain the profile.

• None of the fatty acid profiles provide a good explanation for the fatty acid profile of human milk. Surprisingly, the fatty acid profile for the "3 sweet fruits + avocado" may be slightly "closer" to human milk than goat milk.

The standard fruitarian "party line" regarding the bogus fruit = milk theory is that it applies to sweet fruit only. That is, avocados are not included in the theory. Further, the "party line" is that avocados are a concentrated food, to be eaten sparingly. (In reality, they are a staple food for many rawists, including fruitarians.)

Avocados were included in the analysis here because as soon as you debunk one theory, the promoters of crank science theories are usually quick to make new, unfounded claims. That is, debunk the milk = sweet fruit theory, and it becomes a new "sweet fruit + avocado" theory. Including avocados in the analysis here prevents that from happening.

Also, the analysis of avocados given here uses basically one particular analysis of avocado: in effect, one data set. The nutritional composition of avocados varies substantially, depending on:

• Race of avocado--Mexican, Guatemalan, West Indian, or hybrid

• For a given variety of avocado, the composition varies by season (see Slater et al. for comparison tables).

Finally, fruitarians may try to exploit the result shown above which indicates that the fatty acid profile for "3 sweet fruits + avocado" may be "closer" to human milk than is the goat milk profile. The section above prevents that--the fatty acid profile is just one aspect of many in a detailed analysis. Also, the discussion above concerning the "insulin-killer" sugar in avocados raises real questions regarding the merits of a diet of only sweet fruit and avocado.

GO TO NEXT PART OF ARTICLE

(Discussion: Probing Fruitarian Defenses of the "Fruit is Like Mother's Milk" Theory)

See Table of Contents for Section III - Challenging Fruitarian Defenses of the Theory

Back to Waking Up from the Fruitarian Dreamtime
Back to Research-Based Appraisals of Alternative Diet Lore