The "walk through" count we do each Fall is to determine the approximate percentage of Brevard's squirrels that are of the white variant.  It is designed to prevent the same squirrel from being counted twice but it does not account for those squirrels that are not observed at all.  All squirrels can not be expected to be active during the two hour period of the count, let alone be seen by human observers.  Thus, the count undoubtedly underestimates the squirrel population.  But by how much.  Several current studies are being undertaken to establish a more accurate census.  These fall under two categories (1) repeated mark and recapture in which the ratio of "new" to "previously  captured" specimens caught in traps estimates the number of "unseen" squirrels, and (2) distance mapping techniques which use estimates of near-by (easy to see) squirrels to estimate the number of far away (therefore, harder to see squirrels) to measure the so-called coefficient of detectability.  Once completed and if in agreement, we will know the actual number of squirrels in certain regions; by comparing this to the numbers seen during the October count, we hope to determine a ratio of Obervsed/Actual and be able to extrapolate the total number of squirrels, of both gray and white variants, in Brevard.  However, these procedures are time consuming and we do not anticipate having such good estimates for a few years (preliminary results are discussed below).

Meanwhile, we accidentally made a very simple observation in the Fall of '98 which allowed us to determine the MINIMAL underestimation of squirrels in a two hour count of a particular sector (Sector 13).  On October 29th, Ben Berry and Scott Galloway conducted 4 consecutive "counts" beginning at 7:30 a.m., 10:30 a.m., 1:00 p.m., and 4:05 p.m., respectively.  The purpose was to determine peak squirrel activity periods.  The conventional wisdom is that squirrels are most active in the hours after sunrise and before sunset but our personal experience with the Fall count seemed otherwise.  The results were as follows::
 

Count	White Squirrels	Gray Squirrels	Total	% White
7:30	2	13	15	13.3
10:30	3	6	9	33.3
1:00	0	2	2	0
4:05	5	9	14	35.7

Overall, they made 40 sightings of which 10 (25%) were white.   Squirrel activity level appeared to be negatively correlated with human activity on campus (a second team was sent out to repeat the procedure in a residential area were human activity from 10-2 would be minimum due to school and work but they did not complete their observations).   Without any attempt to mark or otherwise identify squirrels, it would seem that there are a minimum of 18 squirrels in that Sector since 13 Grays were seen in the early morning count and 5 white in the late afternoon.  The most seen on any one count was 15 or 83.3% of the minimum.  The average number of sightings per count was 10 or 55.6% of the minimum number actually present.  Of course, that assumes that all the other squirrels observed (other than the 13 Gray and 5 White mentioned) were re-sightings and that all squirrels were seen in the course of the four counts, both assumptions that are probably incorrect.  Thus, it seems reasonable to suspect that the actual number of squirrels in Brevard is in excess of twice what we ordinarily observe in our Counts.

The number may actually be much greater than that.  Students in the Spring 1999 Animal Ecology course did line transect (aka strip census) estimates of density in six representative Sectors.  This procedure involves measuring (using the focusing mechanism of a telephoto camera lens) the right angle distance of sighted squirrels from several 100 meter transects walked by the observer.  Certain assumptions regarding the relationship of the probability of detection with distance from the transect line allow one to estimate the actual density of squirrels in the vicinity.  These preliminary estimates are in the range of 2.3 squirrels per acre.  The entire study area has in excess of 1,280 acres (2+ square miles) which might then be currently supporting approximately 3,000 squirrels.  This year (Fall 1998) the most we observed on any given Count was 353 or about 11.8% of that estimate.  The most we've observed on a single Saturday morning since the count began was 735 or 24.5% of that number.  Of course, this latter figure was for Fall 1997 while the density estimates were made in the Spring of 1999.  Judging from the Fall 1998 Count data, numbers are down by about 50% from Fall 1997 (following a bad mast year for almost all tree species).   If  the real density is actually down by half of the previous year, then 735 squirrels would represent 12.25% of the actual population, in the same ballpark as this years rate of detection.  In other words, it looks like on any given Saturday during the Fall Count, we see about one eighth of the actual number out there.  These estimates are very crude and sketchy.  Extrapolating to the whole study area is not justified at this time but as our methods improve and our observations accumulate, we should be able to make more definitive statements regarding the actual number of squirrels in Brevard.  For right now we will have to be satisfied with the determination that whatever the absolute number, approximately 20-25% are of the white variant.   Members of the Animal Ecology group include Jeremy Benjamin, Krista Benken, Lauren Eversoll, Bob Glesener (hey, that's me), Brian "Word" Houston, Matt Moreland, and Katie Smith.

The densities below were determined using the strip line transect method which involves walking the designated length in such a manner as to not allow overlap in observable areas and recording not just the number of squirrels observed by their distance from the transect line.  Distance can determined using a 35 mm camera telephoto lens to focus on the squirrel and reading the distance from the lens barrel.  More recently a laser rangefinder has been utilized.  Perpendicular distances were used eliminating the need to determine angle of observation.  However, this meant having to associate each squirrel observed with a landmark in its immediate vicinity allowing for squirrel movement before the observer is able to reach the perpendicular point on the transect line; when many squirrels are observed at one time (as is often the case on the Brevard College campus), this is a formidable task.
 
 

Date	Sector	Beginning Time	Transect Length (m)	# Squirrels Observed	Density (squirrels/acre)
Spring 2001	25/30		660.4	10 (9G/1W)	4.299
20 June 2002	13	10:30	715.35	15 (9G/6W)	1.576
22 June 2002	03	16::00	966	2 (1G/1W)	0.155
24 June 2002	03	11:50	966	10 (8G/2W)	0.652
26 June 2002	24	12:00	1126.54	6 (6G)	0.435
27 June 2002	13	~11:30	715.35	16 (11G/5W)	3.591
28 June 2002	03	~11:00	966	5 (1G/4W)	0.493
24 June 2003	13	13:30	715.35	14 (9G/5W)	2.428
29 May 2004*	13	10:00	715.35	12 (8G/4W)	4.5
29 May 2004*	13	15:00	715.35	11 (8G/3W)	4.44
29 May 2004*	13	18:00	715.35	14 (8G/6W)	4.64
30 May 2004*	13	15:00	715.35	11 (5G/6W)	1.14
30 May 2004*	13	18:88	715.35	11 (6G/5W)	3.54
* Part of White Squirrel  Festival: transect line baited with field corn for 3 days prior to tours.					Average = 3.65