PIT-tagged lizards were released in a series of experiments designed to
assess short-term and long-term movement patterns.
The frequency and extent of legless lizard movement among microhabitats
was previously unknown. Heterogeneity
in vegetation indicates below-ground variation, such as root structures and prey
availability. Soil also provides
habitat heterogeneity; highly compacted soil may preclude lizard burrowing (Hunt
1997), and moist soil may attract lizards (Bury and Balgooyen 1976, Fusari
1984). Because lizards are
ectothermic, we expected that behavioral thermoregulation affected the
activities of legless lizards (e.g. vertical movement), although perhaps not as
much as in their non-fossorial counterparts.
The specific heat of moist sand is greater than air and thus moist sand
insulates legless lizards. In
addition, these lizards are capable of maintaining their standard metabolic rate
at low temperatures (Brattstrom 1965, Withers 1981, Fusari 1984).
This may allow Anniella to actively forage during colder night
hours and throughout seasonal temperature changes (Fusari 1984).
Legless lizards occur at 5.2 to 31.2 degrees C, with 40
degrees C being lethal
(Miller 1944, Gorman 1957, Fusari 1984).
In
prior studies, the fossorial nature of Anniella made it difficult to
determine the effects of increased disturbance to coastal dune habitats, and to
estimate home-range and dispersal ability.
The Moss Landing site, with its sandy soil and varied vegetation cover,
provided suitable conditions to test these aspects of the life history of
legless lizards. Hunt (1984)
completed an exhaustive study of Anniella and found no external sexual
dimorphism, nor are there any known behavioral differences.
Because of this, and because all tagged lizards were adults, we did not
test for gender or ontogenic differences.
To
detect microhabitat selection and responses to disturbance and other lizards,
short-term surveys (24-48 hours) were conducted, which consisted of releasing
animals under various environmental conditions.
Long-term (24 months) monitoring involved releasing a large number of
lizards throughout the 2.43 hectare study site, and tracking them on a weekly to
monthly basis. This portion of the
study focused on the broader goals of estimating home range and dispersal
ability. These studies were
undertaken to (i) test the hypothesis that legless lizards moved in directed
ways within microhabitats, (ii)
evaluate lizards responses to disturbance,
(iii) estimate home-range area, and (iv) determine dispersal ability.
Lizards
were provided food and water until their release. Individual PIT-tagged legless
lizards were randomly selected and transported to the field in plastic
containers containing soil. They
were released at a starting location, and tracked with the PIT-tag reader.
We placed small plastic bags filled with sand and marked with the
lizard’s unique number on the soil’s surface above lizard positions each
time a lizard was detected. The lizards were not captured. The reader operator
systematically searched an area by extending the tag reader in front, thereby
avoiding walking on any un-searched ground.
To avoid foot-fall disturbance, all other measurements were made after
the operator searched the area. We conducted more directed searches as we
approached a previous position of a lizard, moving the reader in concentric
circles (a 1 to 2 m2 area) radiating from the bag.
In addition, as each experiment progressed, we tested an increasingly
larger area outside the perimeter of the release area.
Distance,
from its previous position or from a central location, to the nearest 5 cm, and
the bearing to the nearest 5o was recorded for each lizard
position. Because of the size of
the location marker and rounding procedures, lizard positions within 20 cm of a
previous position were recorded as the same geographic location (i.e. no
movement). Positional data were
converted to geographic coordinates using MATLAB (1998).
Experiments were designed to test 3 parameters: distance, time, and
turning angle for each step in a movement path.
Sampling periods were spaced 4 to 6 hours apart to avoid over-sampling
and to ensure that each step in a path was independent and not autocorrelated.
We
measured site fidelity by comparing the actual movements of lizards with 100
random paths in a Monte Carlo simulation. To
create random paths, randomly generated angles were produced and combined with
randomly chosen (without replacement) distances between successive steps in the
actual path (Spencer et al. 1990). Mean
squared distance (MSD) measured the dispersion around the animal’s center of
activity, with low MSD indicating greater site fidelity.
An animal’s movements were considered non-random when the MSD of the
actual path was less than 95% of the MSD’s for the randomly generated paths.
We produced actual and random MSD’s using
the Animal Movement Analyst Extention (AMAE) program for ArcView (Hooge
and Eichenlaub 1997).
Only animals with 3 or more recorded positions were used in analyses.
All other statistical analyses were done using Systat 9.0 (1999) unless
otherwise noted.
We released 20 PIT-tagged lizards in each of three microhabitats at
random distances and bearings within a 15-m diameter experimental area.
This area contained 1) yellow lupine bushes, 2) grass/forbs, and 3) open
sand. The grass/forbs category
specifically included California poppy, mugwort, phacelia, creeping wild rye,
oats, annual fescue, ripgut brome, wire lettuce, and sow thistle. The 60 lizards were released on 26 June 1998 between 1630 h
and 1700 h; the first sampling period started at 1800 h. Every 4 hours a search of the area was conducted during a 48
hour period after release. As each
lizard was located, the position was marked and time, date, soil temperature
(surface, at 2.5 cm, and 10 cm depth), and habitat data (bush, grass/forbs, or
sand) recorded (Fig. 10). Air temperature was
recorded at the start of each period. We
measured soil compaction at 8 cm and 15 cm depth using a Soil Compaction Tester
(SCT; Dickey-John Corporation).
This instrument measures the
pounds/square inch (psi) of pressure exerted as it is pushed downward into the
soil, and has markings on the steel rod indicating the depth being measured.
To avoid injuring buried lizards, we slowly inserted the thermometer and
SCT probes into the soil ~1 to 1.5
cm away from their known location. Distance
and bearing were recorded from the lizard’s previous position.
Nine temperature monitoring stations were positioned within the
experimental area, three placed randomly in each of the three microhabitats.
Following the procedures outlined above, soil temperature data were
collected during each sampling period; the mean temperatures from these stations
were compared with the mean temperatures at the sites where lizards were
detected. After ensuring that
sample variances were equal with an F test (Zar 1984), we used a two-sample
t-test to compare the means for each habitat type.
We log (X+1) transformed data with unequal variances.
We used a subset of lizards (n = 34) for statistical analysis of
microhabitat and movement patterns. Only
data from lizards detected more than 2 times after release were used in
analyzing site fidelity. Six other
extremely sedentary lizards were excluded because the analysis required
independence of position between steps. Movement
between microhabitats was assessed by comparing the original microhabitat at
release with the habitat each lizard eventually moved into by the end of the 48
hours.
We tested the effects of soil moisture on habitat selection by producing
wet and dry areas and by comparing the movement patterns of lizards between
them. To moisten the soil to a
depth of ~10 cm, 1.3 cm diameter irrigation drip line was used 24 hours before
the start of the experiment. We
constructed an array consisting of a 12.5-m long irrigation hose with six
additional 3-m long irrigation segments running at 90o on alternating
sides of it (3 each side, Fig. 11). Three
randomly chosen “dry” segments were fitted with plugs, so that no water
flowed from them. The center hose supplied water to the “wet” segments, but
emitted no water. Low, dry grass
covered the ground under the array. On
22 July 1998, we released 8 lizards, simultaneously, at the center of each
segment. We searched for the 48
lizards every 6 hours for 48 hours.
To test the effect of lizard density on movement, we tracked lizards in
small groups of 8 lizards and large groups of 20 lizards.
Three sets of each group were released in similar habitat no less than 5
m apart. Beginning 22 July 1998, we
monitored the movements of these 84 lizards every 6 hours for 48 hours.
Between 18 June and 03 July 1998, we released groups of 10 to 12
PIT-tagged lizards at 23 randomly chosen, widely-spaced locations.
To test for home range and long-term movement patterns, we monitored, for
24 months, the activities of these 238 animals, and all free-roaming lizards
released in controlled experiments (478 total lizards).
Because large areas of the site were covered in dense bushes and berry
thickets, only about 45-55% of the 2.43 hectare habitat was accessible. The
amount of accessible area varied seasonally.
Workers searched the accessible areas weekly from July through November
1998, bi-weekly from December through September 1999, and monthly thereafter.
Searches were directed, i.e. we purposely searched on days with higher
temperatures and during the particular time of day we thought we would find the
maximum number of lizards.
Positions
(DGPS) of lizards were recorded with a Trimble ProXr unit and, after April 1999,
were post-processed as described previously. Time of day, weather, soil moisture, and habitat (sand,
grass, lupine, etc.) were recorded. The
lizards and the habitat were disturbed as little as possible.
Lizards that stayed in the same position for more than two consecutive
sampling periods were excavated to test whether the signal was coming from a
loose PIT-tag.
We
calculated the home range for individual lizards meeting three criteria: (i) the
animal was successfully tracked for >180 days, (ii) at least four geographic
positions were obtained, and (iii) the lizard did not die or leave behind a
loose tag during the sampling period.
Two
home range estimators, kernel home range (KHR) and minimum convex polygon (MCP)
were calculated using the Animal Movement Analyst Extention (AMAE) program for
ArcView (Hooge and Eichenlaub 1997). We calculated the 95% and
50% utilization distributions (UD) for kernel home range using fixed kernels
with a smoothing factor calculated with the ad hoc value, as suggested by Hooge and
Eichenlaub (1997) and Worton (1989).
The 95% UD is the area an animal actively uses, whereas the 50% UD
establishes the core area of activity for each lizard.
Because spontaneous tag loss was low in the lab, we assumed that loose
PIT-tags found in the field indicated the death of the animal. We also assumed that all loose tags were found.
The finite survival rate (FSR) was calculated for lizards released in the
field between 18 June 1998 and 08 September 1998, and for tagged lizards held as
controls in the lab (Krebs 1999).
We monitored 38 randomly placed 60 x 60 cm wood coverboards for 21 months (November 1998 through September 2000) within the restored lizard habitat (6 acres). Vegetation was removed so that the coverboard could be set firmly on the ground. Monitoring consisted of scanning each coverboard with the PIT-tag reader before lifting it, then quickly digging through the sand underneath to a depth of ~15 cm. Lizards found underneath the boards were checked for PIT-tags.
We installed and monitored 25 pitfall traps within the collection site, prior to clearing the area of lizards. Pitfall traps consisted of 1-gallon plastic buckets placed so that the lip of the buckets were ~5-8 cm below ground surface. Square plywood tops, raised on 1.5 cm tall legs, provided cover from direct exposure to sunlight without preventing access for small animals traveling along the ground surface. Drain holes were drilled in the bottom and along the side of each bucket, and a 5-6 cm thick layer of sand was added to each bucket to reduce the risk of predation on trapped legless lizards, by small mammals or other reptiles. Pitfall traps were searched on 6 days of each week (30 days) over a 35-day period.
We established 10 plots ranging in size from 213 to 1773 m2 within the collection site. The position and dimension of plots were determined primarily by the requirements of construction activities at the site, therefore they range broadly in size and types of habitat included. Along the perimeter of each plot we dug an approximately 15 cm wide trench, to a depth of ~55-60 cm. The trenches were then lined with an impenetrable barrier constructed with 62.5 cm tall black tar roofing paper, sunk to a depth of ~55 cm. Wooden stakes placed at 1.5 m intervals along the tar paper barrier supported the above ground portion. Soil removed was replaced and compacted on both sides of the barrier. This barrier was established to contain the population of legless lizards within each plot and to prevent legless lizards from re-entering plots once they had been removed.
Timed fixed-area sampling was conducted by 4 to 6 searchers limited by time to 30 minutes. Timed fixed-area surveys were conducted on consecutive days within a 3 to 4 day period. Sampling was conducted during daytime hours between 0900 h and 1800 h. All legless lizards captured during timed fixed-area surveys were retained for the period of the survey and were then replaced at the location from which they were collected. Timed fixed-area sampling requires a more intensive strategy, which may result in more extensive disturbance of habitats in order to maximize sample size. With this in mind, we tested timed fixed-area sampling within each plot while varying the impact of the survey method on the habitat.
Searchers conducting low-impact timed fixed-area surveys were instructed to search for legless lizards, within the plot boundary, both on the surface, under dried vegetation or objects which might provide cover, to a depth of 5 to 7 cm below the surface, while minimizing disturbance of dominant perennial vegetation and annual vegetation, and their associated microhabitats (e.g. duff layers). Searchers were instructed to search small sections of the habitat at a time and to restore any disturbed materials, to as close as possible, to pre-sampling conditions. No tools (rakes, shovels, trowels, etc.) were used during low-impact timed fixed-area surveys.
Moderate-impact timed fixed-area surveys involved more extensive disturbance of vegetation and of duff layers below the larger plants and patches of debris. Searchers were instructed to search for legless lizards to a depth of ~15 cm below the surface by removing patches of annual vegetation and by pushing aside but not uprooting larger perennial plants.
Random quadrat surveys were conducted in each plot by placing a 0.25 m2 square quadrat at randomly pre-selected coordinates. After documenting plants species, vegetation within the quadrat was rapidly removed along with soil to a depth of ~15 cm, to search for legless lizards on or below the surface. Ten to 40 quadrats were sampled within each plot, depending on the plot size.
Following the surveys, vegetation from each entire plot was removed by hand and all legless lizards found were captured. Each plot was then systematically raked 4 to 5 times by 12 to 15 people and all legless lizards found were captured and housed temporarily or relocated into areas not affected by construction. Plots were considered depleted of legless lizards when 1 or fewer legless lizards where found per 40 searcher hours.
To test possible upper carrying capacities in varied habitat conditions, six 3 m x 5 m enclosure structures were installed in the field. Three were placed in high quality habitat within the 6-acre release site on Moss Landing Hill, and 3 were installed on the dune at the site of the former Moss Landing Marine Laboratories. This second site is inferior habitat with immature vegetation. It is comprised of shallow, coarse, low nutrient sand. Low-impact surveys in 1998 and 1999 revealed only a few legless lizards there.
To minimize habitat differences within each site, enclosures were placed near each other. The perimeter of each enclosure was trenched and then lined with an impenetrable barrier constructed with 62.5 cm tall shade cloth, sunk to a depth of ~45 cm. Wooden stakes placed at 0.5 m intervals along barrier support the above ground portion. Removed soil was replaced and compacted on both sides of the barrier. This barrier was established to contain the existing population of legless lizards within each enclosure and to prevent other legless lizards from entering.
In June 1998, PIT-tagged lizards were released within the enclosures. The enclosures are 15m2, and under the very best conditions, would have 25 lizards in them (at 1.67/m2), but more than likely have a maximum of 15. By adding 15 lizards, we at least doubled the population in one enclosure. Thirty lizards were added to a second enclosure, potentially tripling the number of lizards within it. A third enclosure has no added lizards and acts as a control. This procedure was duplicated for the 3 enclosures on the lower dune at the old lab site.
We searched each enclosure for tagged lizards monthly for at least 15 minutes (excluding processing time) on two occasions. Protocols for semi-annually excavating lizards and documenting health were as described above for long-term monitoring of free-roaming lizards; enclosures were extensively searched on eight occasions.
To determine the long-term effects of population density and habitat quality on legless lizards, the enclosures will be excavated in June 2003. We will compare the number, weights, and SVL of animals with control enclosures.
All
lizards quickly burrowed upon release and were completely buried within
approximately 10 seconds. At no
time were healthy lizards, tagged or untagged, seen above ground.
We also noticed that lizards found near lupines and other bushes were
more frequently located at the drip-line around the outer perimeter of the bush
vs. the interior areas of the canopy. The
maximum observed speed for these lizards was 1.96 m/hr (4.9 m in 2.5 hours).
It
took approximately 2 hours to complete searches during each sampling period for
the study of microhabitats. Forty-eight
of the 60 lizards (80%) were detected with the reader at least once in 48 hours. Of those found, 53% were located 3 times or more with a
maximum of 9 times (Fig.12).
Lizards were located in almost equal numbers on the first and second
days.
Thirty-two
percent of the lizards stayed within the same habitat as where they were
released. Fifty percent of lizards stayed in or traveled into bushes, 38% into
grass, but only 12 % for sand (Fig. 13). Eleven
lizards were sedentary, remaining in the same location (from 2 to 9 times) for
multiple sampling periods. The
maximum time a lizard was stationary was 43 hours.
Twenty-seven of the 34 lizards exhibited random movement.
Twenty-one percent of the lizards originating in bushes showed site
fidelity, as did 33% for grass, but only 9% for those in sand.
Lizards
actively moved during day and night; sedentary behavior occurred in daytime as
often as night (Fig. 14). Almost 48%
of the lizards were detected between the hours of 1400 h and 1800 h (Fig.
15),
which coincided with the greatest average soil temperatures at sites where
lizards were found (Fig. 15b).
Maximum
number of detections of lizards occurred at about 15-20 degrees C for all soil depths
(Fig. 16). Lizards were found
within 11.5 cm of the surface at all times of the day as air temperatures ranged
from 11 to 25 degrees C. The surface of
the soil where lizards were found had a much greater range of temperatures than
the air, apparently gaining and then retaining heat.
Temperature extremes were mediated at 2.5 cm depth, showing less
variation that at the surface (13 to 34 degrees C), and at 10 cm, temperatures were
even more homogenous (10 to 32 degrees C). Lizards
were routinely found when the full range of temperatures from all depths near
the lizard were below 20 degrees C (Appendix 1).
Under
bushes, temperature (surface and 2.5 cm depth) where legless lizards were found
was greater than controls. Lizards
were found at positions under bushes where mean temperature at the surface was
20.05 degrees C (standard dev = 4.77), significantly greater than 18.00 degrees
C (standard
deviation = 4.40) at the control stations
(p = 0.04). This same pattern held for lizard detections at a depth of 2.5 cm (mean
= 19.97
degrees C, standard dev. = 4.25) when compared with the control
stations (mean = 18.24, standard
deviation = 2.71; p = 0.03). There
was no statistical difference in the means between lizard locations and control
stations under bushes at 10 cm. For
grass/forb and sand habitats, there were no significant differences in mean soil
temperatures where lizards were found compared with the control stations (Table
7).
Loose soil (0 to 50 psi) occurred at 8-cm depth around the perimeter of
yellow lupine bushes and some forbs (mugwort in particular, Fig.
17).
In some cases, this phenomenon persisted to a depth of 15 cm.
Deeper soils were generally more solid, and areas of open sand were very
compact, up to 160 psi at 15 cm.
The
number of lizards detected (48, 100% of those released) at wet release sites was
equal to the number detected at dry release sites.
More than half of the animals (26) stayed near their original release
habitat (Fig.18). Eight lizards
moved from their original habitat into wet areas, but five moved from wet
habitat into dry habitat. Eight
additional lizards moved into lupine bushes just outside the experimental area.
Data
for 34 animals met the requirements for site-fidelity analysis; twenty-five of
them (74.5%) exhibited non-random movement.
Lizards released in dry areas displayed site fidelity more often (76.9%)
than those released in wet areas (50%).
Twenty-three of 24 lizards released in low-density groups were detected
at least twice (95.8%). More than
28% (17 of 60) of the lizards released in high density groups quickly dispersed
out of range, either vertically or horizontally, and were never found during the
experimental period. Of those that
were located, they were found less often compared with those released in low
density groups.
Data for 37 animals met the requirements for site-fidelity analysis comparing density. Lizards released in low density groups displayed site fidelity 37.5% of the time vs. 33.3% of those released in high density groups. However, only a third of the high density lizards were located often enough to be used in this analysis.
Individual
lizards were located 868 times during 2 years of long-term monitoring.
Four-hundred and fifteen of the 478 (86.8%) free-roaming lizards were
found at least once, and new animals continue to be found that have not been
located since their release. Two
animals were recently found whose last known position was recorded nearly 2
years ago. Fifty-seven of the
animals were located on 4 or more occasions and one was found 12 times (Fig.
19). There was no apparent seasonal pattern in the number of lizards found per
hour of effort, except that in March 2000 102 lizards were located, many of them
appearing gravid. A greater
percentage of legless lizards under bushes were found near the drip line (71%)
vs. the near the roots (3%) or in the interior of the canopy (26%).
Forty-one
lizards, tracked for 243 to 671 days, were used in home range analyses (Appendix
2). The number of geographic
positions recorded per individual was 4 to 12, and the maximum distance
traversed was 34.8 m in 305 days (11.4 cm/day).
We noted no homing tendency (i.e. no lizards were observed making
straight-line paths toward their original capture location).
For
the kernel home range analysis, the mean 95 % UD estimate was 71.0 m2
(standard deviation = 87.2), and the mean 50% UD was 15.8 m2
(standard deviation = 21.3). The
large variances in the home range estimates were attributed to 5 extremely
sedentary lizards (95% UD < 10 m2) and 3 lizards with 95% UD’s
>225 m2 (Figs. 20, 21,
22).
The mean size of home range UD’s did not increase with the length of
time lizards were tracked or sample size (Fig.
23).
The MCP home range was 0.33 to 70.96 m2 (mean = 13.30 m2, standard deviation = 16.97).
During
summer 1998, 568 tagged lizards were released.
Between 30 June 1998 and 30 June 1999, 12 loose tags were found (FSR for
year 1 = 97.9%). An additional 26
loose tags were located the following year (FSR for year 2 = 95.4%, overall FSR
= 93.3%). Only five tags had animal
tissue still attached to them, so approximating the date of death was not
practical. Of 530 lizards that
presumably survived through 30 June 2000, 44% of them (232) were located within
the prior six months. The FSR for
tagged control lizards held in the lab was 90.4% for year 1, 84.2% for year 2,
with an overall FSR of 80.9%.
An
owl pellet containing a PIT-tag from a legless lizard was found 800 m away from
the lizard’s release site, and a second loose tag was found a similar distance
from its lizard’s origin. Five
tags were found in scats of Felis catus.
A robin (Turdus migratorious) captured but did not kill 2 legless
lizards, a hawk was seen flying with a live legless lizard in its beak, and a
partially digested lizard was found in a bolus, presumably from a marsh hawk (Circus
cyaneus).
Lizards were rarely found under coverboards, except during the mating season in April and May 2000 (Table 8). The estimated population of lizards in the 24,000 m2 study area is 9,000 (estimated at the mean density of the cleared area, 0.228 lizards/m2, plus ~3,500 recovered lizards). This density was validated by the mark-recapture estimate of 9,360 lizards. Given these estimates, 0.0067 percent of the total population was detected using coverboards. Of the 478 free-roaming tagged lizards, 0.0063 percent were detected using coverboards.
We found 3 legless lizards in 25 pitfall traps in 35 days. Two animals were found in one of 12 pitfall traps placed in a plot which supported a relatively low density of ~0.08 legless lizards/m2. We found one animal in 1 of 13 pitfall traps in a plot which supported at least 701 legless lizards (a relatively high density of ~0.40 animals/m2).
Moderate-impact timed fixed-area searches consistently revealed more animals than the other survey methods. On average, moderate-impact timed fixed-area surveys revealed twice as many legless lizards (mean = 9.6, standard dev. = 4.9) than were found with low-impact timed fixed-area surveys (mean = 4.9, standard dev. = 2.4), a statistically significant difference (T9 = 4.10, P = 0.002). At very low densities (0.01 animal/m2) no legless lizards were detected with either timed fixed-area sampling method. At slightly higher densities (0.08 to 0.12 animals/m2), we found 5 to 9 animals with moderate-impact area searches and 3 to 7 animals with the low-impact method. In plots supporting the highest estimated population densities (0.40 to 0.58 individuals/m2), we found 7 to 13 animals with moderate impact sampling while 3 to 8 were found with the low-impact method. We found no correlation between the number of legless lizards detected during timed fixed-area surveys and either the minimum size of the population within each plot or the estimated population density.
The area covered by randomly placed 0.25 m2 quadrat included from 0.3% - 2.0% of each plot. Random quadrat surveys did not reveal the presence of legless lizards in 5 of the 10 plots, which supported the lowest and highest minimum population densities estimated (0.01 and 0.58 individuals/m2). In the remaining 5 plots, where population densities ranged from 0.11 to 0.54 legless lizards/m2, we detected 1 to 5 animals. Sampling with 0.25 m2 quadrats plot revealed from 2 to12 plant species in each plot. In general, plots with higher plant diversity supported higher legless lizard densities, yet this pattern was not exclusive. In plots with relatively high species diversity (11 to 12 plant species), legless lizard densities ranged from 0.12 to 0.54 individuals/m2 (mean = 0.36). In plots with plant species diversity ranging from 4 to 7, legless lizard densities ranged from 0.08 to 0.58 individuals /m2 (mean = 0.31). In the 2 plots with the lowest plant species diversity (2 species), legless lizard densities were 0.01 and 0.10 individuals/m2.
One hundred percent of tagged lizards released in enclosures were detected with the reader at least once, 42% of them in the previous 6 months. We continue to detect lizards previously “missing” during extensive bi-annual searching, indicating that lizards remain out of range of the reader; the enclosures contain significant vegetation cover that reduces the searchable area. No lizards were found outside the enclosures and no free-roaming tagged lizards have been found inside the enclosures.
At least three lizards were preyed upon. These loose tags were transported (one was found in an owl pellet) from 20 to 800 m away from the release location. Three other loose tags were found. There is no clear indication that lizards from any one enclosure are missing in higher proportion or experienced greater mortality (Table 9). Most the re-weighed animals from the high quality habitat enclosures gained weight (10 of 12, + 0.4 to 0.6 g). Two lizards from each of the low quality habitat enclosures were re-weighed. In both cases, one animal gained weight (+ 0.2, + 0.5g) while one lost weight (-0.3, -0.3 g).
All
experimental sites supported an existing population of un-tagged legless
lizards, which may have affected results. We
assumed that lizards acted independently, that is the effects of other lizards
on results were not known but likely equal among test groups.
All experiments were conducted at one geographic location with presumably
high habitat value.
Searching
for lizards may have resulted in changes in lizard movement patterns.
We attempted to test the effects of human foot-fall disturbance by
removing ground vibration caused by walking. We simultaneously released 8
lizards in the center of a 9 m2 area.
The area was bounded by a “catwalk” (Fig.
24) made of four 3m-long wood boards
supported by four hay bales. Movements
of lizards were monitored from atop the catwalk without entering the area on
foot. We repeated the experiment twice with inconclusive results.
On
the first occasion, lizards were released in an area vegetated with small yellow
lupine bushes. We tracked their movements every 15 minutes for 24 hours, and
found that although the lizards moved often, they did not move far, remaining
under or around the bushes (Fig. 25). This
could have indicated that lizards move less after eliminating vibration from
foot traffic. On the second
occasion, the area contained only grass and forbs.
Within 12 hours, 7 of 8 lizards made long movements out of the area below
a large blackberry (Rubus ursinus) patch to the southwest (Fig. 26). The
overall conclusion is that lizards seek and settle into quality habitat.
Effects of foot-fall disturbance, however, remains undetermined.
In
all of the short-term experiments, we carefully transported animals, allowed
them to acclimate at the site within their individual containers, and handled
them as little as possible. Initial
movements, however, were probably enhanced in response to their handling and
disturbance (Turchin 1998).
Individuals
undetected during experiments were an indication that (i) the lizard had moved
into an un-searchable area, (ii) it had burrowed deeper than 10 to 11.5 cm, or
(iii) the lizard was near the surface, but was in a vertical position such that
the tag did not reflect the reader signal.
These limitations created a bias toward animals that remained near the
surface or those that remained in a more horizontal position and reduced overall
sample size and increased the variability in the number of detections per
lizard.
We
did not collect detected lizards, and so did not determine their depth in the
soil. Therefore, soil temperature
was only known to within 10 cm of the actual position of the lizard, and soil
compaction is only known to within 15 cm.
Because
only adult lizards were large enough to tag, the movement patterns of younger
lizards was not tested. Because
lizards were not sexually dimorphic, we could not test differences between male
and female activities. Unless a
lizard remained in the same location more than twice, we assumed that the lizard
was alive.
The
high percentage of lizards relocated was encouraging, and because we continue to
find lizards that have not been located for long periods, we should be able to
refine home range estimates through time. Although
some lizards may frequent shallower burrows more often, the recapture
probability was not biased toward slow or easily captured animals.
Anniella
is almost exclusively fossorial, therefore, we assumed that all movements
occurred underground. In the field,
animals were never seen on the surface and laboratory observations, including
time-lapse video, support this finding. Lizards
probably used the soil interface for feeding and for mating when there was
sufficient leaf litter to conceal them.
Our
observations also indicated that lizards routinely occupied deeper soil than was
previously reported. “Missing” lizards often re-appeared in the same spot,
indicating that they migrated deeper and then returned to the shallower
location. Although results may be
an artifact of the terrariums used, lab observations clearly indicated that
lizards were capable of burrowing to at least 46 cm depth.
Soil compaction readings in the field indicated that Anniella can
move through relatively compact sand. A
review of the literature revealed several ideas about the way that these lizards
move within their subterranean environment. Miller (1944) mentioned burrow systems, and Fusari (1984) and
Kamel and Gatten (1983) identified Anniella as a sand swimmer, indicating
that soil re-consolidates behind the animal as it moves underground.
Our lab observations confirm that legless lizards can build elaborate,
persistent burrow systems. These
burrows, and the low soil compaction that indicated their presence, may provide
a new tool in the difficult task of detecting the presence of Anniella.
This burrowing may play an important role in the ecosystem by
increasing soil aeration and drainage. The
influence of burrows between roots could be important to plants and their
subterranean insect fauna (Maron 1998).
Lizards
actively moved day and night in all of the short-term tracking studies.
A high percentage of lizards were found between 1400 h and 1800 h when
soil temperatures were greatest, indicating that lizards used the temperature
gradient of the soil to thermoregulate. Anniella
does not avoid temperatures below 20 degrees C, as indicated by the thermal
gradients used by Bury and Balgooyen (1976).
Although we found lizards in soil 27 degrees C, we were unable to establish an
actual upper temperature limit. This
is because lizards were not collected and we did not know the exact temperature
at the exact depth each was found. In
selecting areas under bushes with higher soil temperatures, lizards may take
advantage of the sunnier sides of the vegetation for basking.
Short-term movements may not be a response to immediate needs for
resources such as food or water. Animals
were fed and well hydrated when they were released. Overall, nearly half of the
animals used in the test of short-term site fidelity had random movement paths.
Anniella pulchra has a relatively low standard metabolic rate,
54-81% of the mean for other reptiles of the same size (Kamel
and Gatten 1983), and may not feed often (Hunt pers. comm.).
Still, lizards released in bushes tended to stay there, whereas lizards
released in sand eventually traversed into grass or bushes. Many of the lizards released in grass also moved under yellow
lupines where arthropod prey density may be greater. Lizards had no immediate need to find soil moisture; animals
released in dry areas did not necessarily seek soil moisture horizontally nor
did they vertically migrate to moisture below the surface.
The
extent to which lizards avoid each other is unknown.
Lizards held in the laboratory were normally well-spaced, and there were
no signs of aggression among them. When
placed in greater densities, they dispersed more than lizards released at low
density.
We
observed that lizards tended to select the outer perimeter of the canopy of
bushes. More of the lizards found
under bushes were located at the drip line and low soil compaction readings at
these areas indicated a large number of burrows.
The drip line may provide more moisture than surrounding sand, greater
soil temperature, or there may be other unseen differences at this ecotone.
Home
range is usually described as the area an individual uses for feeding, mating,
retreat, basking, and other normal activities and excludes occasional sallies to
explore new habitat (Burt 1943). Kernel
density estimators, non-parametric approaches to determining home range area,
are considered robust, and are becoming widely used (Worton 1987, Worton 1989,
Seaman and Powell 1996, Hansteen et al. 1997).
Our small sample size probably biased the results by overestimating the
true home range size (Worton 1987, Seaman and Powell 1996). Even given this flaw, the kernel estimator model provides a
plausible home range estimate.
MCP
home-range estimation is an older, commonly used test in early studies of
reptiles. It is very sensitive to
sample size: as sample size increases, so does the size of the home range (Worton
1987, Boulanger and White 1990, White and Garrot 1990). This explains the small mean MCP of 13.30 m2.
MCP home range was determined by drawing a polygon which encompasses all
of the known positions for an animal and calculating the total area within it.
Utilization distributions are more effective for estimating the habitat
animals use because they are probabilistic; each known position has an
“associated probability that the animal is in that location” (Hooge
and Eichenlaub 1997).
Home
range sizes varied greatly among individual lizards in this study. It is possible that
some of the variation in home range size was attributed to differences in
ranging distances between genders. This
is a common phenomenon in terrestrial lizards.
Turner et al. (1969) summarized home ranges for 14 different
insectivorous lizards, and found that females had much smaller ranges (15 to
1000 m2) than males (10,000 to 20,000 m2).
Some
lizards moved 10 to 25 m within the first month after release before settling
into a smaller area. This may be the time it takes for lizards to readjust to
their surroundings after the disturbance of being moved.
The small mean home range for this population of A. pulchra (71.0
m2, 95% UD) may be due to a high abundance of food, soil moisture,
and other required resources.
Animals
found moving into higher quality habitat may be seeking less patchily
distributed resources. Anniella feeds on beetle larvae, adult beetles, insect
pupae, spiders, sow bugs, ants, and termites, many of which are probably
abundant on Moss Landing Hill (Coe and Kunkel 1906, Miller 1944, L. Kuhnz, pers.
observ., Fusari pers. comm.). Hunt
(1984) stated that food was probably not a limiting factor in the size of
legless lizards populations. Home
range size can change through time due to resource availability and population
density, so we will continue to monitor this population of A. pulchra to
gain insights into the status of the relocated lizards.
Although
no consistent pattern of seasonal activity has emerged, the large pulse of
apparently gravid lizards found during March indicated that perhaps reproduction
activity can be monitored in the future.
There
are no data on longevity for A. pulchra.
This lizard exhibits low fecundity (1-4 live-born young) and has a low
metabolic rate, so it may be a long-lived species (Miller 1944, Goldberg 1985,
L. Kuhnz, pers. observ.). Lizards
can live as long as 7.5 years under laboratory conditions (Krieberg pers.
comm.). The presence of loose tags
in the field was probably a good indication of mortality because spontaneous tag
loss was low under laboratory conditions. The
mortality rate, however, may be underestimated.
Ailing and dying lizards have a strong tendency to come to the soil
surface (L. Kuhnz, pers. observ., Krieberg pers. comm.).
This means that the tags left behind after the carcass of a lizard
decomposes are very likely to be near the soil surface where they can be found
with the reader. However, mortality
may be greater than we can calculate because predators carry animals, with their
tags, to inaccessible places. Mortality
in the laboratory was greater than field mortality, and may indicate not all
dead animals were identified in the field.
Domestic
cats are a problem in developed areas and readily dig for lizards (Hunt and
Zander 1997, L. Kuhnz, pers. observ.). Other
likely predators present on the site include 6 to 8 local raptors, deer mice,
skunks, opossums, gray fox, red fox, weasels, coyotes, and dogs.
Determining
home range and long-term dispersal ability for a population is an important step
toward resource management and for planning future translocation projects.
Relocation
methods should be thoughtfully considered.
Workers should exercise care when releasing animals of indistinguishable
gender and small home range into large areas.
Inadvertently releasing large numbers of all male or all female animals
at widely spaced intervals may cause a population decline as animals travel
greater than normal distances to find mates.
Timed fixed-area surveys proved to be the most reliable method for detecting legless lizards over a wide range of population densities, vegetation types, and over a relatively wide range of areas surveyed. The results of these search methods however failed to correlate with our estimates of the minimum population size or population densities. Timed fixed-area searches also failed to reveal the presence of legless lizards at very low population densities (~0.01 animal /m2). Although we typically found twice as many legless lizards with moderate-impact timed fixed-area searches relative to the low-impact timed method, low-impact timed fixed-area searches proved to be as equally effective at revealing the presence of legless lizards.
The instructions we gave to searchers conducting low-impact timed fixed-area searches, were to minimize disturbance to the vegetation and the surface of soil and duff layers, so that visual evidence of the search would for the most part be erased by the next heavy rain. While this is a highly qualitative parameter, inspection of the plots indicated that resulting disturbance to the habitat was substantially less during low-impact searches than during moderate-impact searches, or during our sampling with randomly placed quadrats.
The failure of timed fixed-area searches to successfully correlate with either the minimum population size or the minimum density of legless lizard is most likely the result of large variations in plot size, the range of legless lizard population densities, the distribution of plant communities within each plot, and the variable experience of searchers, all factors which differed between plots. Timed fixed-area searches conducted by 4 -5 people over an area no greater than ~1000 m2 may be the most productive. Searches conducted within smaller areas revealed high numbers of animals, yet resulted in more concentrated disturbance of the habitat. Searches in a very large plot revealed relatively low numbers of animals, given the high population density (~0.40 legless lizards /m2). Additional refinements of this survey method will likely enhance its demonstrated use in evaluating the characteristics of legless lizard populations.
Although
the percentage of the minimum estimated population within each plot, revealed by
pitfall trap arrays was similar to the overall percentage of animals detected
with low-impact timed fixed-area sampling, obtaining results with these passive
methods required an extensive sampling period (> 30 days). Similarly low
returns have been reported by Block et al. (1988) who detected 1 legless lizard
in pitfall traps located in an oak woodland at Tejon Ranch, Kern County
California during 7848 trap nights. We were unable to gather sufficient data
from pitfall trap arrays to predict a minimum period over which these methods
could be employed with negative results before an area is declared devoid of
legless lizards. Coverboards are a
poor indicator of population density and should not be used as an exclusive
means to determine that legless lizards are absent from an area.
Coverboards did not become “seasoned” through time, i.e. they did not
become more effective the longer they remained in place.
We feel strongly that the wide-spread use of this method as a means for
surveying for legless lizards is highly misleading.
Our results provide strong evidence that coverboard transects do not
provide sufficient return over limited periods of time to accurately determine
presence or absence of this species. Extensive searching, even if moderate
habitat destruction occurs, is the only viable way of ensuring that an area is
devoid of legless lizards. Burrows,
and the low soil compaction that indicates their presence, may provide a new
tool in the difficult task of detecting the presence of Anniella.
We
will actively track free-roaming and enclosed lizards for three more years,
including monitoring their general health once every 6 to 12 months.
Longer term monitoring will provide more well-defined home ranges, and
information about carrying capacity and longevity.
Advances
in the power of the electromagnetic field generated by PIT-tag readers will soon
allow us to find lizards even deeper underground (up to 21 cm).
This will provide an opportunity to gain information about vertical
migrations in Anniella.
We
hope to develop a lavage procedure for obtaining stomach contents from live
legless lizards. Stomach content
analysis along with food availability studies in the field will help us more
fully understand what constitutes quality habitat for legless lizards.
Anniella
is closely related to anguid lizards,
some of which exhibit highly developed chemosensory abilities (Cooper 1989,
Cooper 1990, Cooper 1992). Testing
legless lizards for these abilities may provide insights into how they locate
food and mates. For example, their
ability to detect yellow lupine bushes may lead them to rich food resources.
Management
of this species is dependant on the ability to find better ways of determining
population density and long-term stability of those populations.
Exporting these methods to new locations will allow us to determine if
home range size is related to the local setting (i.e. rich dune soil, and a
restored, dynamic habitat) or if other microhabitats support lizards which
utilize space differently.