Honors Thesis Proposal

download Honors Thesis Proposal

of 21

Transcript of Honors Thesis Proposal

  • 8/8/2019 Honors Thesis Proposal

    1/21

    Variation in pocket gopher (Thomomys) digging morphology

    as a determining factor in species distribution innortheastern California

    Ariel E. Marcy

    Objective of research

    Pocket gophers (family Geomyidae) are subterranean rodents that use tooth- and claw-

    digging to varying degrees to tunnel underground. Western pocket gophers (genus

    Thomomys) are distributed throughout the western U.S. Thomomys inhabit a wide variety

    of habitats and exhibit huge interspecific morphological variation. Pocket gopher ranges

    are determined through competitive exclusion such that two species rarely occupy the same

    area.

    My objective is to combine species locality and morphology analyses to test

    whether variations in the Thomomys digging apparatus can explain why each species gains

    competitive dominance in the particular soil and climate conditions of its range.

    AIM 1: I hypothesize that morphological variation between the digging apparatus

    of the species determines their niche along the tooth-to-claw-digging strategy spectrum. I

    will test this using morphometrics quantitative size and shape analyses to describe and

    compare gopher skeletal specimens from all five species in NE California.

    AIM 2: I predict that these differences contribute to the competitive dominance of

    one species over another under certain soil characteristics (e.g. particle size, hardness,

    water capacity), many of which vary with climate. I will test this by using Geographic

    Information Systems (GIS) to project the latitude and longitude of gopher capture site data

    onto maps of soil and climate characteristics across California.

  • 8/8/2019 Honors Thesis Proposal

    2/21

    AIM 3: Use GIS to analyze the smaller subset of morphometrically-described

    specimens to directly explore the relationship between individual morphology and the soil

    and climate conditions of its habitat. If strong correlations exist, I will use them to propose

    a biomechanical explanation of how gopher morphology determines each species

    competitive ability under certain soil and climate characteristics. This knowledge could

    help explain the current distribution pattern of pocket gophers across California,

    understand a species turnover event during the Pleistocene, and predict gopher

    distributions under a climate change model.

    Introduction

    Charles Thaeler laid the groundwork for pocket gopher research in Northeastern California

    by investigating the claim that gopher ranges never overlap. Thaeler chose Northeastern

    California because five species inhabit this area and of the ten possible pair-wise species

    overlap combinations, nine occur. He created a distribution map for NE California through

    extensive field observations and sampling (Fig. 1). He determined that Thomomys have

    abutting ranges, but species almost never exist sympatrically (Thaeler 1968). The rare

    cases of range overlap occur in areas with two radically different soil types. Morphological

    differences had not been analyzed thoroughly by Thaelers time but he suggested these as

    the main reason for competitive exclusion (Thaeler 1968).

    Thaelers hypothesis for exclusion assumes that variations in environmental

    characteristics across the NE Californian region alter the outcome of competition. To test

    this, my study will combine species locality and morphology analyses to see whether a

    biomechanical advantage under particular soil and climate conditions can explain the

    geography of range boundaries between competing gopher species. Extensive

  • 8/8/2019 Honors Thesis Proposal

    3/21

    documentation of gopher sensitivity to soil moisture, temperature, and texture supports this

    hypothesis (Miller and Gross 1998, Romanach et al 2005). Digging is very energetically

    demanding, but the exact cost varies between soil types (Lessa 1990). Furthermore,

    metabolic rates of subterranean species working in harder soils vary, demonstrating that

    some are more efficient in particular soil types (Luna and Antinuchi 2006).

    Recent assessment of gopher morphology has emphasized the dual nature of the

    digging apparatus, involving both the forelimbs and the jaw, which are the lever arms for

    claw- and tooth-digging respectively (Lessa and Thaeler 1989). Further analysis of the

    digging apparatus found that claw-digging tends to restrict animals to friable sandy soils

    whereas tooth specializations broaden the potential spectrum of habitable soil textures

    (Lessa and Thaeler 1989). The authors conclude that digging specializations contribute to

    the realized range of gophers, however, the exact mechanisms are undeveloped.

    In a follow-up study, Lessa and Stein conducted linear morphometric and

    myological analyses that included two representatives from the genus Thomomys. The

    morphometric analysis found only osteological variation between jaw and forelimbs (Lessa

    and Stein 1992, Stein 1993). This observation allows me to limit my morphological studies

    to skull and forelimb bones, which are much more accessible than myological material.

    Their analyses also showed that Thomomys morphology varies from highly specialized for

    tooth-digging to slightly favoring claw-digging (Fig. 2).

    Despite the large number of morphological studies on pocket gophers, no studies

    have focused only on differences within a genus and thus range determination between

    species is still poorly understood (E.A. Hadly, personal communication 2009). The

    existing literature draws on the larger cross-genera morphological differences to

    characterize gophers as more specialized for claw-digging versus tooth-digging

  • 8/8/2019 Honors Thesis Proposal

    4/21

    invaluable information for my project. However, unwarranted generalizations from one

    species to its entire genera are common in pocket gopher literature. Despite the spread in

    Thomomys digging modes noted in Lessa and Stein 1992, the genus is usually

    characterized as tooth-specialized diggers in cross-genera studies, and strangely, Stein

    describes them as claw-diggers in Stein 2000.

    Morphological comparisons within any pocket gopher genera would generate

    interesting results to a field lacking such studies. Thomomys is an ideal genus to investigate

    sub-genera differences because its species vary across an order of magnitude in size and

    they inhabit a wide range of soils and habitats (Stein 2000, Thaeler 1968). Most of the

    cross-genera studies described above used linear morphometrics. My approach will

    achieve a higher degree of sensitivity by adding geometric morphometric analysis.

    My focus on interspecific differences allows me to introduce a unique extension to

    gopher ecology. Previous cross-genera studies made observations of the general ecology

    for the genera, but extensions are limited because each genus of gophers resides in

    different regions of the country. Therefore the populations studied do not interact and

    comparisons cannot control for larger variations between ecosystems or for variations

    within the genera. I avoid these issues by focusing on all the species within one genus

    located specifically in Northeastern California.

    If both facets of my study reveal significant results, these will be combined to

    propose a biomechanical explanation for the competitive exclusion determining pocket

    gopher spatial distributions. Knowledge of why particular gopher species competitively

    exclude the others in certain soil types and that of ancient gopher distribution information

    could be extended to recreate ancient regional climates. There is evidence of a gopher

    replacement event in Northeastern California during the Pleistocene in which one species

  • 8/8/2019 Honors Thesis Proposal

    5/21

    supplanted another from its former range (Blois 2008). We will be able to describe how

    climatic conditions might have changed during the Pleistocene, once we understand the

    soil and climate conditions these species thrive in.

    Furthermore, this knowledge, in conjunction with climate change models, could be

    used to predict future gopher distributions under different parameters. Gophers are

    ecosystem engineers that create habitats for endangered species, such as the burrowing

    owl, reduce invasive grass establishment, and otherwise greatly affect community structure

    (Cameron 2000, Cox and Hunt 1994, Eviner and Chapin 2003, Stein 2000). Forming a

    deeper understanding of pocket gopher species biomechanics, competitive interactions, and

    ecology will generate insights into the past, current, and future Californian ecosystem.

    Materials and Methods

    First, I will use two types of morphometric analysis, linear and geometric, to

    quantify functional differences between the skulls and forelimbs of the five species of

    Thomomys pocket gophers from NE California. Second, I will use GIS to overlay museum

    specimen locality data on maps of soil and climate. Principal Components Analyses (PCA)

    and Geographically Weighted Regression (GWR) can determine whether correlations exist

    between the soil, climate, and the spatial distribution patterns exist across the western US,

    concentrating in California. Finally, I will analyze the subset of gopher species from NE

    California with morphometric data in GIS to determine whether the morphological

    differences can be linked to differences in the abiotic characteristics where they occur to

    help explain the trends seen across species in the Western US.

    1. Analysis of variations in Thomomys digging morphology

  • 8/8/2019 Honors Thesis Proposal

    6/21

    Morphometrics is a quantitative way of comparing shape. Linear morphometrics

    (LM) are simple measures of length and width of bones and bony processes (muscle

    attachment sites). I will use five forelimb as well as eight cranial and dental standard

    measurements previously described for gophers (Fig. 3). My specific predictions for

    variations in linear measurements by digging mode are described in Table 1. LM can be

    used on ancient bones and will generate absolute measurements appropriate for

    quantitative biomechanical studies of gopher lever arms. Often morphometric studies on

    species within a genus produce data that can be used to identify ancient bones more

    specifically.

    Geometric Morphometrics (GM) is a branch of mathematical shape analysis aided

    by digital camera and computer software. GM can perform more sophisticated analyses

    than LM because linear measurements do not convey information about spatial

    relationships between measurement points (Zeldich et al 2004). GM solves this problem by

    creating a digital network of points called landmarks that define the shape of the bone of

    interest. Landmarks must be discrete points on an organism that are recognizable across all

    specimens and can be pinpointed with a computer image of the specimen (Fig. 6). Often,

    recognizable areas have functional significance, such as a muscle attachment scar. I can

    use the photos I take for LM analyses and ImageJ to perform most GM analyses. The

    Hadly lab computer has additional statistical analysis programs for regression, CVA,

    MANOVA, and PCA tests.

    I will continue to acquire gopher skeletal specimens from the Museum of

    Vertebrate Zoology (MVZ) at University of California, Berkeley. Only female specimens

    will be used for morphometric studies because their growth ends short after reproductive

    maturity, unlike males, whose continued growth could introduce allometric errors related

  • 8/8/2019 Honors Thesis Proposal

    7/21

    to specimens age at capture (J.L. Patton, personal communication 2010). Only

    individuals collected from Northeastern counties will be used in order to control for

    regional variation across the state. Furthermore, because gopher populations exhibit huge

    intraspecific variation, I will break down species and subspecies into 14 geographic groups

    after Patton and Smith 1990. I aim to measure at least 25 specimens per geographic group,

    which is the precedent from previous morphometric studies in rodents (Anderson et al

    2002, Cartini and OHiggins 2004, Zeldich et al 2004, Zelditch and Swiderski 2009).

    Because GM organizes landmarks into a matrix of interrelated points, each

    specimen can be scaled to the same size, thus removing the variable that usually drives

    most of the differences in the linear PCA and can overshadow variation due to shape. A

    PCA of GM data creates vectors that indicate the direction of shape change across groups

    of species (Fig. 7). Landmarks could then be grouped using osteological or myological

    adaptation indicators of digging mode previously described by cross-genera studies of

    pocket gopher morphology. Analysis of how these elements vary could then be applied to

    describe how the NE Californian Thomomys geographic groups fall with respect to each

    other on the claw- and tooth-digging spectrum.

    2. Geographic Information System (GIS) analyses of Thomomys species of California

    I will use the Geographic Information System (GIS) to test whether range

    boundaries of species are correlated with particular soil and climate conditions by

    projecting the latitude and longitude of each gopher capture site onto maps of soil and

    climate. I will compare species locality data to soil clay, sand, and silt content, water

    capacity, clay mineral type, rock strata, temperature, elevation and annual precipitation

    maps. I can access precise maps of soil and climate characteristics for the Western US

  • 8/8/2019 Honors Thesis Proposal

    8/21

    from the Natural Resources Conservation Service (NRCS) and United States Geological

    Service (USGS).

    Gopher locality data for all five species ofThomomys from museums in the western

    region of the US can be downloaded from both the Mammal Networked Information

    System (MaNIS) and Arctos, a multi-institution museum database. The dataset of gopher

    localities will be reduced to just those points with georeferencing error smaller than 3km.

    These data will be used to generate maps illustrating patterns observed across the entire

    western US concentrating in California but including Oregon and Nevada. This region is

    environmentally heterogeneous and will give a large area to observe patterns over.

    Results can be analyzed on an individual map basis by graphing the soil or climate

    parameter versus species on density histograms and comparing the averages between

    geographic groups. These analyses can be done for the entire geographic range or for much

    smaller regions. Data can be exported from GIS and the analysis and graphing can be done

    in R. All of the results for every climate and soil parameter will be analyzed in a more

    powerful Geographically Weighted Regression (GWR). Using the GWR algorithm in GIS

    will determine which abiotic parameters are involved in determining gopher ranges and

    which are the most important. GWR is particularly useful because it weighs each

    parameters importance over the entire range, so the analysis will give maps showing how

    the influence of each parameter varies over the landscape.

    3. GIS analysis of morphometrically described specimens in NE California

    To determine whether the patterns observed at the species level are in fact driven by

    morphological differences, I will use GIS to analyze the 14 groups of morphologically

    described specimens from NE California. NE California is an ideal study area because of

  • 8/8/2019 Honors Thesis Proposal

    9/21

    the ten possible pair-wise species interactions, nine occur. Analyses used for the

    geographic groups above, can be applied to individual linear measurements or trends in

    geometric morphometrics. To reduce error in my analyses, I will need to return to the field

    notes for each of the specimens I choose to ensure more accurate georeferencing. When

    running GIS analyses, it is possible to introduce a buffer zone that takes into account the

    relatively large extent gophers inhabit and the error introduced when field researchers take

    only a few georeference points for each trapline.

    Preliminary and Anticipated Results

    I became involved in the Hadly lab the summer after freshman year working as a field

    assistant to Jessica Blois, a graduate student in the lab at the time. She was investigating

    how climatic change since the last ice age to the present had affected small mammal

    populations. When I began working in the lab my sophomore year, I met with Prof. Hadly

    to discuss potential projects I could work on. Prof. Hadly suggested Thomomys as an

    interesting study species to describe morphologically because there was some debate over

    an ancient species found only in the Shasta cave where Jessica had been working. Over the

    course of the year, Jessicas results revealed an interesting turnover event in which one

    species of gopher pushed another north. These results, my interest in multidisciplinary

    studies, and more discussions with Prof. Hadly changed the focus of my project to the

    interaction of biomechanics and environment in range determination over space and time.

    1. Analysis of variations in Thomomys digging morphology

    Using digital calipers, I have completed forelimb measurements for 98 individuals

    including all five species. Redundant measuring allowed me to assess whether my

  • 8/8/2019 Honors Thesis Proposal

    10/21

    measurement error was below 5%. I was not able to measure one specific measurement

    consistently with calipers so I plan to take digital pictures of my specimens and use ImageJ

    to retake them. Using R, a statistical program, I have completed a preliminary PCA

    analysis of the five species by the four accurate measurements of their forelimbs (Figs. 4,

    5). PCA is a mathematical procedure that transforms a number of possibly correlated

    variables into a smaller number of ranked uncorrelated variables called principal

    components (Zeldich et al 2004). In morphometrics, the first PC usually describes size and

    the succeeding PCs account for shape. The results show separation of species by size and

    the correlation of PC2 to the deltoid process suggest that more powerful analyses like

    geometric morphometrics will find significant shape differences.

    2. GIS analyses of Thomomys species across California

    With the help of Hadly lab technician, Lily Li, and Branner GIS specialist, Patricia

    Carbajales, I have completed substantial analyses of soil clay content and species locality

    (Figs. 8-10). Using data from NRCS soil surveys of the United States, we grouped the soil

    survey units into bins of 0-20%, 20-40%, and >40% clay content as suggested by visiting

    physical soil properties specialist, Professor Oliver Chadwick (Figs. 8, 9). Overlaying

    species locality data over the maps appeared to indicate that the claw-digging subgenera

    Thomomys species tracked the regions with lower clay content whereas the tooth-digging

    subgenera Megascapheus species tracked regions with higher clay content (Figs. 8, 9).

    Density histograms of species from small geographic regions showed that they are indeed

    sorting by clay content (Fig. 12).

    Unexplained portions of the map, however, include the Northeast corner of

    California, where a subgenus Thomomys species, T. mazama, inhabits a region of high soil

  • 8/8/2019 Honors Thesis Proposal

    11/21

    clay content. I hope to continue investigating how soil characteristics interact with climatic

    factors to produce these types of results.

    3. GIS analysis of morphometrically described specimens in NE California

    I have found preliminary results using the linear morphometric data points for the

    forelimbs of 30 female individuals with less than 3km georeferencing error. Correlations

    between soil clay content and forelimb measurements were all positively correlated with p

    values less than 0.01. All of my predictions for this relationship were met except for

    humerus length, which increased (Table 1, Fig. 13). I had expected this to decrease due to

    increased forces experienced by the forearm while digging (Stein 2000). However, it may

    have increased as a result of size conferring an advantage in hard soils. Ulna length also

    increased with a correlation greater than that of olecranon length, which also suggests size

    playing a role.

    These results did not break the points down by species, so some of the results, such

    as increased ulna and humerus length, could be related to the species inhabiting harder

    soils, emphasizing tooth-digging over claw-digging. To investigate these trends by species,

    I will need to gather a larger sample size for each group, which will be accomplished for

    the 14 geographic groups of 25 individuals from NE California. Importantly, these results

    show that correlations between soil type and species from part 2 are seen also among

    individuals, which supports a biomechanical explanation for range determination as

    opposed to shared characteristics of a species.

    The preliminary results seem promising, however, much more data needs to be

    collected and analyzed more thoroughly before conclusions can be made.

  • 8/8/2019 Honors Thesis Proposal

    12/21

    Appendix

    Fig. 1: Distribution map of Thomomys genus pocket gophers, for reference, top border is Oregon,

    right border is Nevada (from Thaeler 1968)

    Fig. 2: PCA illustrating how genera and two species of Thomomys spread on axes of size and

    claw- or tooth-digging traits (from Lessa and Stein 1992)

  • 8/8/2019 Honors Thesis Proposal

    13/21

    Fig. 3: Measurements for morphometric analysis (from Lessa and Stein 1992)

    Table 1: Linear morphometrics measurements, abbreviations, grouped by forelimb andskull measurements, Hypotheses given for morphology changes in harder soil. Preliminary

    results available for forelimb measurements (n=30).Measurement Abbr. Claw-digging Preliminary Results

    Humerus

    length

    HUML Decreased (higher forces experienced

    in arms)*

    Increased (cor 0.59, p

  • 8/8/2019 Honors Thesis Proposal

    14/21

    Fig. 4: PCA of my preliminary linear measurements. PC1 is correlated with size.

    Fig. 5: Scatter plot of P21 versus deltoid width shows that PC2 is moderately correlated to

    variation in the deltoid process.

  • 8/8/2019 Honors Thesis Proposal

    15/21

    Fig. 6: Landmarks for a South American subterranean rodent, Ctenomys torquatus skull (From

    Fernandes et al 2007)

    Fig. 7: PCA results of landmark-based GM of a rodent cranium (From Zelditch and Swiderski

    2009)

  • 8/8/2019 Honors Thesis Proposal

    16/21

    Fig. 8: Illustrator-generated image of Fig. 1 over the soil clay content map of California

    Fig. 9: GIS-generated map of species point localities over California region clay content,demonstrating more robustly the patterns of subgenera tracking with low and high % clay

  • 8/8/2019 Honors Thesis Proposal

    17/21

    Fig. 10: GIS-generated map of NE California and the proposed 14 geographic groups formorphometric analysis (part 1) and GIS analysis of morphometrically described

    individuals (part 3)

  • 8/8/2019 Honors Thesis Proposal

    18/21

    Fig. 11: Topographic map of NE California groups with Mount Shasta study area

    highlighted

    Fig. 12: Density histogram of 3 species near Mount Shasta showing separation by % clay

  • 8/8/2019 Honors Thesis Proposal

    19/21

    Fig. 13: Scatter plot showing correlation (0.59, p

  • 8/8/2019 Honors Thesis Proposal

    20/21

    References

    Anderson, R. P., Peterson, A. T. and Gomez-Laverde, M. 2002. Using niche-based GISmodeling to test geographic predictions of competitive exclusion and competitive

    release in South American pocket mice. Oikos 98: 3 16

    Blois, J., 2008. Small mammal response to the Pleistocene-Holocene transition innorthern California. Journal of Vertebrate Paleontology 28, 53A -54A.

    Cameron, G.N., 2000. Community Ecology of Suberranean Rodents. In: Lacey E.A.,

    Patton J.L., Cameron G.N., (Eds.), Life Underground. The University of ChicagoPress, Chicago, pp 227-256.

    Cartini, A. and OHiggins. 2004. Patterns of morphological evolution in Marmota

    (Rodentia, Sciuridae): geometric morphometrics of the cranium in the context ofmarmot phylogeny, ecology and conservation. Biological Journal of the Linnean

    Society, 82, 385407.

    Cox, G.W. and Hunt, J., 1994. Pocket gopher herbivory and mortality of Ocotillo onStream Terrace, Bajada, and hillside sites in the Colorado Desert, Southern California.

    The Southwestern Naturalist 39, 364-370.

    Eviner, V.T. and Chapin, F.S., 2003. Gopher-plant-fungal interactions affectestablishment of an invasive grass. Ecological Society of America 84, 120-128.

    Fernandes, F.A., Fornel, R., Cordiero-Estrela, P. and Freitas, T.R.O. 2009. Intra- and

    interspecific skull variation in two sister species of the subterranean rodent genusCtenomys (Rodentia, Ctenomyidae): coupling geometric morphometrics and

    chromosomal polymorphism. Zoological Journal of the Linnean Society 155, 220-237.

    Lessa, Enrique P., and Barbara R. Stein., 1992. "Morphological constraints in the digging

    apparatus of pocket gophers." Biological Journal of the Linnaen Society 47, 439-53.

    Lessa, Enrique P., and Charles S. Thaeler, Jr., 1989. "A reassessment of morphologicalspecializations for digging pocket gophers." Journal of Mammology 70, 689-700.

    Lessa, Enrique P., 1990. "Morphological Evolution of Subterranean Mammals: Integrating

    Structural, Functional, and Ecological Perspectives." In Liss A.R. (Ed.), Evolution ofSubterranean Mammals at the Organismal and Molecular Levels pp 211-30.

    Luna, F. and Antinuchi, C.D., 2006. Cost of foraging in the subterranean rodent

    Ctenomys talarum: effect of soil hardness. Canadian Journal of Zoology 84, 661-667.

    Hadly, E.A. personal communication 2009

    Miller, M.A. and Gross, M.M., 1998. Locomotor advantages of Neandertal skeletalmorphology at the knee and ankle. Journal of Biomechanics 31, 355-361.

  • 8/8/2019 Honors Thesis Proposal

    21/21

    Patton, J.L, and Smith, M.F. 1990. The Evolutionary Dynamics of the Pocket Gopher

    Thomomys bottae, with Emphasis on California Populations. University of CaliforniaPress, Berkeley.

    Romanach et al, 2005. Effects of species, sex, age, and habitat on geometry of pocketgopher foraging tunnels. Journal of Mammalogy 86, 750-756.

    Stein, B.R., 1993. Comparitive Hind Limb Morphology in Geomyine and ThomomyinePocket Gophers. Journal of Mammalogy, 74, 86-94.

    Stein, B.R., 2000. Morphology of Subterranean Rodents. In: Lacey E.A., Patton J.L.,

    Cameron G.N., (Eds.), Life Underground. The University of Chicago Press, Chicago,pp 20-61.

    Thaeler, Charles S. Jr., 1968. An Analysis Of The Distribution Of Pocket Gopher Species

    In Northeastern California. PhD. Thesis, University of California Press, Berkeley andLos Angeles.

    Weijs WA, Hillen B. 1985. Cross-sectional areas and estimated intrinsic strength of the

    human jaw muscles. Acta Morphol. Neerlando-Scand 23, 26774.

    Zeldich M.L, Swiderski, H.D.S., and Fink, W.L., 2004. Geometric Morphometrics forBiologists: APrimer. Elsevier, San Diego.

    Zelditch, M.L., and Swiderski, H.D.S., 2009. 2009 Berkeley Geometric Morphometrics

    Workshop. Berkeley, California.