The measurements presented in this section are based on 37 points. I made these measurements in late 1996 for Dr. Vance Haynes. In 1973, Dr. James Judge also measured and reported on most of these points. Since I had both sets of data, I have decided to share them with the reader. My measurements were made using Haynes' scheme (Figure 1) and therefore are slightly different from Judge's. These differences are explained throughout this section. Figure 1 belongs to Dr. Haynes and I am displaying it with his permission. I have modified it slightly for quicker loading.


Haynes' scheme requires a basal width measurement to be made with the point and the calipers against a base line (flat surface). The point is positioned with its longitudinal axis perpendicular to the base line. If the point flares out from the proximal edge then the basal width at the thickness of the calipers is recorded. If it tapers in, the basal width at the baseline is measured. Therefore, the thickness of the calipers can make a difference, but it is small. (I used a plastic set that were 4 mm thick.) Judge reported basal widths measured at the proximal edge of the point (1973:149).

Table 1 -- BASAL WIDTH and 1/2 BASAL WIDTH

basal width (mm)1/2 basal width (mm)
NMeanSt. Dev.NMeanSt. Dev.

In Table 1, the 1/2 basal width measurement is Haynes' nomenclature. Judge termed the same measurement as "hafting width". The difference in my numbers and Judge's numbers is due more to the additional artifacts I measured than the difference in measurement methods. Two of my base fragments were too short to yield the 1/2 basal width statistic and this is the explanation for the 35 bases in my 1/2 basal width statistic. Graph 1 presents the relationship between base width and 1/2 base width.

Graph 1

The linear regression line in Graph 1 is expressed by 1/2 basal width = 1.01*(basal width) + 0.58. This line explains 93% of the variation in the 1/2 basal width parameter which is large for lithic measurements in my experience. I am accustomed to seeing numbers in the 60 and 70% range. This high percent of correlation is also present between 3/4 basal width and basal width (Graph 2) and discussed later.


3/4 basal width (mm)one basal width (mm)
NMeanSt. Dev.NMeanSt. Dev.

Comparison of Table 1 & 2 indicates that the bases expand toward the tip (distal end) of the point. Additionally, the rate of expansion is very uniform across the assemblage as was evident in Graph 1. Graph 2 continues to support this uniformity.

Graph 2

The linear regression line in the Graph 2 has the equation of 3/4 basal width = 0.94*(basal width) + 2.64. and explains 87% of the variation in the 3/4 basal width parameter. Again, this explained variation is large for this type of data. (I chose not to present a graph for one basal width because it had the same character as Graphs 1 and 2, and would increase the loading time of this page.)


thickness at W3/4 (mm)concavity depth (mm)
NMeanSt. Dev.NMeanSt. Dev.
**Judge's thickness at 1/2 basal width

My measurements of thickness, in Table 3, were recorded at 3/4 basal width while Judge's were made at 1/2 basal width. This is most likely the explanation for the differences in mine and Judge's averages of thickness. Unlike the various measurements of basal width, the correlation between thickness at 3/4W and basal width is nonexistent. This lack of correlation is evident in Graph 3 which is a classic "who flung dung" chart. Here, the linear regression line only explains 9% of the variation in thickness at 3/4W.

Graph 3

The difference in concavity depth in Table 3 is more difficult to explain. I would be incorrect in suggesting the difference was a result of six (6) additional points in my numbers. If this were true then the six (6) additional points would have to have convex bases to account for the difference in magnitude of the numbers. For the record, there are no convex Belen points. If the difference was a result of Judge measuring between ears and I measuring a concavity depth from a flat surface then my concavity depth should be larger than Judge's, which it is not. This was the most difficult measurement for me to make, and I will suggest this is the reason for the difference between our numbers. However, I can not say who is more correct, or if either of us is correct. Graph 4 should give the reader an appreciation of the variation in this parameter. The linear regression line explains only 7% of the variation. Another, classic "who flung dung" plot.

Graph 4

In summary, the base of a Belen point can be described as concave to almost straight. It gently expands toward the tip (distal end) and is approximately 21 millimeters wide and 5 millimeters thick. The widths at various locations along the base are predictable from the basal width, while thickness and basal concavity are unrelated to the basal width.

Judge made a parallel observation. He calculated the coefficient of variation (standard deviation / average) for the measurements he made. His coefficient of variation indicated that the basal width measurements were more consistent than the thickness or concavity depth measurements. He suggested that basal width was related to hafting and thickness and concavity depth were not (1973:216).

My father, in the 1960's, suggested that the plano-convex cross-section of the Belen base was functional and the function was hafting. He suggested the Belen point was socketed inside the rib bone of a large animal as indicated in the drawing on the right. If this was possible and a reality, then this would explain the low variations in the basal width measurements. I have not made measurements on bison or cow rib bones to see if it is plausible.

Return to Paleoindian and Other... Homepage.