Do Voles Select Dense Vegetation for Movement Pathways at the Microhabitat Level?

 

 

Fellow- Lucretia Olson

The University of Montana IBS-CORE Program

Don Christian

 

 

 

Back to Abstract

 

 

Introduction

The meadow vole, Microtus pennsylvanicus, and the montane vole, Microtus montanus, are the main vole species found in the Mission Valley, near Charlo, Montana.  They live in grassy areas, usually with dense thatchy ground cover, and eat mainly grass stems and grains.  Many predators feed on voles, including coyotes, weasels, and raptors (Tamarin, 1985).  Voles play an important role in the ecosystems they occupy since they constitute such a large part of the prey base for many animals.  When voles are scarce, the predators that usually feed on them are forced to turn to other sources of food, such as songbirds and waterfowl eggs and young. 

            The problem of predation hazard and microhabitat has been addressed in a few other studies, notably that of Bowers and Dooley, 1993.  This study determined that voles might select their habitat based on the type of predation that most seriously threatens them, as well as the quality of the patch in which they are living.  It did not address microhabitats within patches, however, but dealt mainly with the effects of habitat edges on vole habitat use.

            My study was designed to determine if there is a relationship between the microhabitat that voles move through and the density of vegetation around the site.  My hypothesis is that voles will actively choose the density of vegetation to move through on a microhabitat level so that they can remain under a maximum amount of cover as a protection against predators.  If this is so, trails made by the voles should display a greater amount of vegetative cover than the area immediately surrounding a trail, which should have a random amount of both dense and sparse cover.  Also, the amount of thatch and bare ground at a location could affect habitat use, since thatch affords good protection, while bare ground should expose the animals to any visually searching predators, and should therefore be avoided.

            Another factor I addressed in this study was whether the tortuosity of a trail varied with the amount of vegetative cover over the trail.  The tortuosity of a trail is a measure of how much twisting and turning a vole does while it is foraging or simply moving through an area.  Theory suggests that animals in a habitat with good cover should move less directly and linearly than those in sparse cover (Stapp, 1997).  This should be because in good cover the voles will feel safe enough to move freely about, foraging, while in sparse cover, the trails will be more straight because the animals are trying to move through a dangerous area as quickly as possible. 

 

Methods

Study Sites

I conducted my study during July and August 2000, in the Mission Valley near Charlo, Montana.  Ten study sites were selected to include a combination of state, federal, and tribal lands.  All the areas were in open, grassy fields, ranging from sparse, dry grasses to more thickly covered fields with tree belts and small potholes.  The sites included areas that had been grazed recently, several years ago, or not at all.  Five of the sites were located on Duck Road, west of Highway 93, 4 were on Highway 12, and one was on west Post Creek Road, 3 miles west of Highway 93.

 

Trapping design and vegetation analysis

I used Sherman live traps baited with rolled oats to capture the voles.  Fifty traps on a site were used, in two rows, with 10 m between traps and 25 m between rows.  I set the traps directly on the ground, moving away vegetation where necessary.  The traps were set in the evening, checked early in the morning, about 12 hours later, and then closed for the rest of the day.  Each site was trapped for a total of three nights.  Captured voles were weighed, sexed, and marked with a permanent marker for recapture.  I then dusted each vole with fluorescent powder by putting then in a bag filled with the powder and gently shaking the bag.  The voles were released at the site where they were captured.

            Vegetation was analyzed at each of the 50 trap stations per site.  Percent cover was estimated using a 0.25 m square and dividing cover into grass, forbs, thatch, and bare areas.  Vegetation height was determined using standard Robel methodology; the height of the vegetation was assessed at each of the four cardinal directions around the trap station and then averaged.

 

Trail evaluation

Trails made by powdered voles were followed in full dark to allow maximum visibility of the trail.  I used a black light to illuminate the powder trail, and followed each trail for 15 m, being careful not to walk on the trail and crush the vegetation.  A roll of surveyor’s tape was laid down along the trail, matching the animal’s movements from side to side and over and under vegetation as closely as possible.  I tried not to disturb any of the vegetation over the trail, and went back along it after tracking to return the vegetation to as near as normal as possible.  The trails were then evaluated the next day.  Total trail length, as well as straight-line distance from start to finish, was recorded.  I used a compass to map the trail and record the directions of any deviations over 0.1 m. 

            Vegetative obstruction was measured every 0.5 m along the trail.  To qualify as open, an area had to have a 0.1 m or larger piece of tape visible through the vegetation at each point every 0.5 m along the length of the trail.  I checked the obstruction from directly above, and 0.5 m to the left, right, front, and back of the trail.  Random visibility measurements of the immediate surroundings were also taken, using a 3 m length of 0.5 in pipe.  At the 3, 6, 9, and 12 m marks on the trail, the thin pipe was slid into the ground cover perpendicular to the trail on either the right or the left side.  I then determined if the area above the pipe was open or covered at 0.5 m intervals using the above criteria.

Data analyses

Vegetative obstruction was totaled for each trail in several ways.  The number of covered points on a trail were counted when looking from straight above the trail, recorded here as “top”, from 0.5 m in front, to the left, right, and above the trail, recorded as “broadview”, from 0.5 m to the front and directly above, recorded as “on trail”, and 0.5 m to the left and right, recorded as “off trail”.  The number of covered points on the “top” view was divided by the total number of covered and uncovered measurements to determine a proportion of obstruction for each trail.  This procedure was followed for the “broadview”, “on trail”, and “off trail” views with the amount of completely open and completely obstructed measurements for each trail.  The proportions for each view, covered and uncovered, were averaged and compared between the measurements taken on the trail and those taken on the transects.  Measurements were compared using a two-tailed, paired student’s t-test.  Average trail segment length was found for each trail by dividing the length of segments by the number of segments in a trail; segments were defined as a straight-line distance between two angles in the trail.  The compass angles of each segment on a trail were combined and used to find a standard error of all angles.  The standard error of the segment compass angles on a trail was used as a measure of tortuosity (i.e. SE=0 for a straight trail).  The standard error was compared with vegetative cover of each trail, and a correlation coefficient  for these two variables was calculated.  Last, the average vegetative height, amount of bare ground, and amount of thatch for each capture location was compared to the number of voles caught at that location, with a correlation coefficient calculated for both of these variables.

           

Results

There was a significant difference between the trail and the random transects for the “top”, “broadview”, and “on trail” views when comparing completely open measurement proportions.  The visibility of trails selected by voles did not differ significantly from the visibility on transects perpendicular to the trail when the completely obstructed proportions of the three views were compared.  Neither the obstructed nor the open measurements were significant for measurements in the “off trail” view.  The proportion of open and obstructed measurements taken every 0.5 m along the trail and those taken every 0.5 m on the transects next to the trail are shown in Table 1, as well as the p-value for each comparison.

            The correlation coefficient between the average segment length for each trail and the proportion of trail completely covered when viewed from the “top”, “broadview”, and “on trail” views are shown in Table 2.  This table also shows the correlation coefficient for these views and the standard error of the compass angles for all segments in each trail.  The average coefficient for segment length is –0.18, not significant at an alpha = 0.05 level, while for standard error of angles it is –0.46, significant at the alpha = 0.05 level. 

            The correlation coefficient between the average heights of the vegetation in each field versus the number of voles caught in that location was 0.1.  For the average amount of thatch in a field versus the vole number, the coefficient was 0.42.  For the average amount of bare ground versus the vole number, the coefficient was –0.34.  None of these coefficients are significant at the alpha = 0.05 level.

 

Discussion 

My results suggest that there is a relationship between the density of vegetation and the areas in which voles travel at a fine microhabitat level.  The comparisons made between obstruction measurements taken on the trail and on random transects were significantly different only when comparing open areas; there was no significant difference in dense vegetation on or off the trails.  These findings indicate that voles are not selecting among dense vegetation, but instead are selectively avoiding sparse vegetation at the microhabitat level.  One possible explanation for this finding is the year this study was done, the local vole population was at or near the very bottom of its cycle.  During this study, I caught only 39 voles, a rate of 2.6 per 100 trap nights, which compares with snap-trap capture rates during earlier studies in this area of 3.0 in summer 1999 and 15.0 in fall 1998.  Thus, voles were at an extremely low density this year.  With fewer voles, there was not an intense social pressure that may force some voles to live in sparse areas.  Other studies have shown that when a population is low or increasing, there is no need for animals to live in sub-optimal habitat, and will occupy preferred areas (Hartman, 1996).  Most of the voles were able to choose where they would live, and so were able to pick areas that afforded them the best protection from predators, those that were in the middle of dense patches of vegetation (Bowers and Dooley, 1993). 

            Vole habitat choice is important because it must take into consideration food and mate availability, shelter, and protection from predators (Stoddart, 1979).  The quality of a microhabitat varies both with forage availability and predator vulnerability.  Most small mammals must reach a balance between the distance they are from cover and the abundance of food when choosing where to move and forage (Lima, 1990).  The apparent lack of distinction by voles on a microhabitat level to select areas that show proportionally more vegetative obstruction should only be observed when the population is low enough to afford choice movement and foraging areas to most of the individuals.  That the voles are significantly avoiding bare areas proves there is enough optimal habitat so each individual can move through only dense areas.  Should this study be performed on a yearly basis through an entire vole population cycle, I would expect the results to reverse, and show a tendency to select denser habitat over sparse, while still avoiding bare places.

            The measure of tortuosity by average segment length in a trail was not significant when compared to the vegetative obstruction in “top”, “broadview”, or “on trail” views.  The standard error of segment compass angles had a significant negative correlation coefficient at the alpha = 0.05 level when compared with “top”, “broadview”, and “on trail” views.  This shows that as a trail gets more obstructed, the tortuosity actually goes down.  Although this is counter to my original hypothesis, it does confirm the previous findings that voles are selecting against bare areas while not distinguishing among dense areas on a fine microhabitat scale.  As the area on and around a trail becomes more covered, the vole may find it less necessary to change directions frequently to maintain protection from predators, and thus the tortuosity would go down.  As the trail area became more open, however, the vole would be forced to switch directions frequently in order to take advantage of what little protection the bare area offered, causing higher tortuosity.  These results are in opposition to the findings of a study done on deer mice, which showed that tortuosity was higher with increased vegetative obstruction, and so should be further explored before any definite conclusions are reached (Stapp, 1997). 

            There was almost no correlation between the number of voles caught in a given location and the average height of vegetation at that location.  This contradicts my hypothesis, in that if the voles were selecting for more dense vegetation, they should be found in areas with higher average vegetation.  Again, however, the lack of a relationship may be attributed to the low numbers of voles present in the study area.  With such a low total population, the numbers of individuals per location does not vary enough to show a strong correlation coefficient.  There was a stronger coefficient, although still not significant, between the amount of thatch and number of voles per location, showing that as thatch increases, the number of voles does so also.  This can be explained by the fact that voles use thatch as a protection from predators, especially aerial predators such as raptors (Sonerud, 1986).  Conversely, there is a fairly strong negative connection between the amount of bare ground and the number of voles at each location.  The voles should avoid bare ground because of its vulnerability to predators. 

All of these facts allow me to conclude that voles selectively avoid sparse or bare vegetation on a fine microhabitat scale.  The proportion of open points on vole trails was significantly lower than on random transects, while the tortuosity for open areas was significantly greater.  This indicates that voles are avoiding bare areas when possible, and moving through them by keeping to as much cover as they can find.  These findings may be reversed when more voles are present in the population, however, and to be conclusive, this study should be performed for at least the length of the vole population cycle, which varies by several years according to many factors such as weather, habitat availability, and food abundance (Chitty, 1996).  Another factor which may prove of interest is whether during high population years voles that have the best habitat are of a certain sex or higher level of fitness than the individuals in the more sparse areas. 

 

Acknowledgments

I thank Don Christian for his help in every step of the project; Carol Olson, Dan Olson, and Steve Olson for assistance during the field work and while analyzing data; Dave Fitzpatrick and                   Joe Quinn for help in the field; and the IBS-CORE grant which made this research possible, funded by the Howard Hughes Medical Institute.

 

Literature Cited

Bowers, M.A., and Dooley, J.L. 1993.  Predation hazard and seed removal by small mammals: microhabitat versus patch scale effects.  Oecologia  94:247-254.

Chitty, D.  1996.  Do lemmings commit suicide?  Oxford University Press, New York.

Hartman, G.  1996.  Habitat selection by European beaver (Castor fiber) colonizing a boreal landscape.  Journal of Zoology, London  240:317-325.

Lima, S.L. and Dill, L.M. 1990.  Behavioral decisions made under risk of predation: a review and prospectus.  Canadian Journal of Zoology  68:619-640.

Sonerud, G.A.  1986.  Effect of snow cover on seasonal changes in diet, habitat, and regional distribution of raptors that prey on small mammals in boreal zones of Fennoscandia.  Holarctic Ecology.  9:33-47.

Stapp, P. and Van Horne, B. 1997.  Response of deer mice (Peromyscus maniculatus) to shrubs in shortgrass prairie: linking small-scale movements and the spatial distribution of individuals.  Ecology 11:644-651.

Stoddart, D.M.  The evolution of habitat selection.  In: Ecology of Small Mammals.  Chapman and Hall, London.

Tamarin, R.H. 1985.  Biology of New World Microtus.  Special Publications No. 8.  American Society of Mammalogists.

 

Executive  Summary

            I studied voles, a small mammal about the size of a large mouse, to determine if there was a relationship between the thickness of vegetation and where the voles chose to move around and eat.  To do this, I live-trapped voles, dusted them in a fluorescent powder, and released them.  I returned at night to illuminate powder trails with a blacklight and mark the paths.  By analyzing the thickness of the vegetation along the trail and comparing it to random samples of vegetation off the trail, I was able to determine if the voles were choosing the areas in which they foraged.  There were slight differences in the vegetation on and off the trail, but statistically, there is no difference between the two.  This allows me to conclude that voles are not selecting for thick vegetation on a small spatial level.

 

PART A.

 

Lucretia Olson

Mentor: Don Christian

Project Title: Do Voles Select Dense Vegetation for Movement Pathways at the Microhabitat Level?

 

  1. I enjoyed my summer research and would say the quality was very high.  The project was interesting and the field work was challenging.

 

  1. Benefits of participation in the IBS-CORE program included experience in field situations, working with professionals in my field, gaining first-hand knowledge on how to put together a project and carry through with all of the stages, and experience in writing and presenting scientific papers.

 

  1. I did not participate in activities in my mentor’s lab.

 

  1. Before beginning work, my mentor showed me the necessary techniques for laying out traps, tracking trails, and analyzing vegetation. 

 

  1. I met with my mentor about every 2 weeks, and he was usually accessible even though he was in Missoula and I was in Ronan.

 

  1. I would recommend Don as a mentor for future students; he was very helpful with the entire project, and also extremely patient in helping me with writing the paper and analyzing the data.  Also, his ideas for research were interesting and good when I was stalled with the early stages of my project.

 

  1. I was fairly well prepared for my research this summer.  I have taken several classes that included designing and implementing a research project, albeit at a small scale.  I felt prepared to carry out the field work due to a summer class with Kerry Foresman, and confident about writing the paper due to classes like Ecology, Animal Behavior, and Developmental Biology.  My statistics knowledge was weak, however, even after I took the required Statistics 241 class, and this has caused me to elect to take Statistics 444 and 445 this year.

 

  1. After I graduate, I plan to go on to earn a Master’s or  a Ph.D in biology.  This fellowship has helped me gain needed experience with skills in my field as well as letting me get to know people who can recommend me to these advanced programs.

 

  1. I probably will not present or publish this work.

 

  1. I have no suggestions, the program seemed well planned.  The summer meetings were helpful and the people were very nice.

 

 

 

 

 

 

 

Back to Abstract