Prabir Burman

Identifying individual mountain lions Felis concolor by their tracks: refinement of an innovative technique

Abstract This study refines a method reported by Smallwood and Fitzhugh (Smallwood, K.S., Fitzhugh, E.L., 1993. A rigorous technique for identifying individual mountain lions Felis concolor by their tracks. Biological Conservation 65, 51–59) that attempted to discriminate between individual mountain lions by certain measurements of their tracks in the field. During the months of January–March 1996, we followed 10 radio-collared mountain lions in the Sierra Nevada of California and obtained photographs of their tracks in the soil and snow under many different environmental conditions. Linear and area measurements were determined from track photographs and Fisher’s discriminant analysis was used to differentiate between each track set. Unlike the Smallwood and Fitzhugh analysis, we were certain about the identity of most of the mountain lions that made tracks. Our results indicate that track sets had both correct and incorrect “groupings” and that these groupings were sensitive to the type of substrate in which a track set was found, the time of day it was photographed, and the number of tracks in a set. In general, it is important to minimize variation associated with substrate and time of day between track sets and to concentrate on sets that contain three or more tracks. This technique has potential application in wildlife conservation; however, the cautionary guidelines, developed in this paper, should be considered.

A data dependent approach to density estimation

SummaryA data dependent approach to density estimation is proposed here. The proposed method requires boundedness and some weak integrability condition on the unknown density, but not any assumption of smoothness. Applications to histogram, kernel, spline and orthogonal series methods are discussed.

Regression function estimation from dependent observations

We consider the problem of estimating a regression function with nonrandom design points and dependent errors. We construct a spline estimate of the regression function and obtain its rate of convergence. It turns out that the dependence of the observations is reflected in this rate.

Multivariate mode hunting: Data analytic tools with measures of significance

Multivariate mode hunting is of increasing practical importance. Only a few such methods exist, however, and there usually is a trade-off between practical feasibility and theoretical justification. In this paper we attempt to do both. We propose a method for locating isolated modes (or better, modal regions) in a multivariate data set without pre-specifying their total number. Information on significance of the findings is provided by means of formal testing for the presence of antimodes. Critical values of the tests are derived from large sample considerations. The method is designed to be computationally feasible in moderate dimensions, and it is complemented by diagnostic plots. Since the null hypothesis under consideration is highly composite the proposed tests involve calibration in order to ensure a correct (asymptotic) level. Our methods are illustrated by application to real data sets.

Estimation of generalized additive models

Spline estimation of generalized additive models is considered here. Cross-validation is used as a criterion of model estimation. Some computationally simpler approximations to cross-validation are given.

Seasonal Patterns of Female Homalodisca coagulata (Say) (Hemiptera: Cicadellidae) Reproductive Physiology in Riverside, California

Abstract Female Homalodisca coagulata (Say) were collected from October 2001 to February 2005 from citrus at the University of California, Riverside. Between 5 and 20 females per sampling date were dissected, and each was assigned an ovarian rank: previtellogenic, vitellogenic, or postvitellogenic. Ovarian ranking was used to characterize H. coagulata reproductive activity. Results of these dissections revealed consistent annual patterns in the proportion of previtellogenic females present in this field population. These patterns indicate that there are two distinct generations annually, with an occasional third generation. A time-dependent model of H. coagulata vitellogenesis cycles in Riverside, CA, was developed, which makes it possible to predict the appearance of the subsequent generation based on previous observed peaks in the proportion of vitellogenic females.

Diet of Florida coyotes in a protected wildland and suburban habitat

Coyotes (Canis latrans) arrived to Florida (USA) in the 1960s and are currently found throughout most of the state. The purpose of this study was to determine if the diet of Florida coyotes differed between suburban and wildland habitat types or across seasons. Fresh coyote fecal samples were collected from wildland and suburban habitats in Pinellas County, Florida (USA; 27°54′N, 82°41′W) from May 2005 to March 2007. Diet items in the 49 wildland and 71 suburban samples were identified to the lowest possible taxonomic level by gross morphological characteristics and medullary configurations of dorsal guard hairs. A Poisson regression was utilized to determine the main and interactive effects of habitat and season on the number of different food items per sample. Coyotes in the wildland habitat had greater diet diversity than suburban coyotes. In addition, anthropogenic waste was recovered over twice as often from coyote fecal samples collected in the suburban habitat. In the wildland habitat, vegetative matter (96%), Insecta (53%), and Rodentia (45%) were recovered most often, as opposed to berries (56%) and Lagomorpha (32%) in the suburban habitat. In both habitats, vegetative matter, berries, and Lagomorpha were recovered most often from coyote fecal samples, whereas Odocoileus virginianus, Lagomorpha, and berries varied the most between wet and dry seasons. This study suggests that as urbanization increases, diet diversity for the coyote will likely decrease and consumption of anthropogenic items will likely increase. As a result of this, human–coyote conflicts may become more common—particularly in counties, like Pinellas, that are undergoing urbanization.

Semiparametric Modeling of Seasonal Time Series

Often, in practice, one may regard an observed time series as being composed of a function that is smooth over years, with additive seasonal effects. As a modification, we formulate a particular multiplicative model that expresses the observed data as a yearly trend function with additive amplitude-modulated seasonal factors. Without smoothness restrictions on the yearly trend and modulation components, the least squares solutions for the seasonal components are shown to be proportional to the eigenvector corresponding to the maximum eigen value of the within-season covariance matrix. If the trend and seasonal modulations are modeled as smooth splines, we give the comparable estimators for the smooth functions and the seasonal factors. We show consistency for the trend, modulation and seasonal factors as well as asymptotic normality for the seasonal estimates. Model selection, fitting and forecasting are considered for a quarterly earnings series that exhibits extreme nonlinear and nonstationary behavior. We compare the results with those obtained using a competing nonstationary multiplicative ARIMA model

A comparative study between enteric parasites of Coyotes in a protected and suburban habitat

Coyotes (Canis latrans) have inhabited Florida (USA) since the 1960s and are currently found throughout the state. Our objective was to obtain information on enteric parasites of coyotes from two different habitats. Fresh coyote fecal samples were collected from protected and suburban habitats in Pinellas County, Florida, USA (27o54′ N, 82o41′W) from May 2005 to March 2007. A standard fecal flotation examination and formalin-ethyl acetate sedimentation protocol were utilized on fecal samples from both habitats. Five newly documented coyote parasites were documented: one cestode (Hymenolepis spp.), one nematode (Ascaris spp.), and three protozoa (Balantidium coli, Blastocystis spp., and Entamoeba histolytica). Nine hitherto unreported parasites for FL coyotes were also discovered: two cestodes (Diphyllobothrium latum and Dipylidium caninum), two nematodes (Toxocara canis and Uncinaria stenocephala), one trematode (Paragonimus spp.), and four protozoa (Cryptosporidium spp., Giardia canis, Cystoisospora spp., and Sarcocystis cruzi). The protected area supported significantly more undocumented (i.e., newly identified) parasites for FL coyotes, and Protozoa as compared to the suburban area. Florida coyotes are likely more susceptible to infection by novel parasites because of their rapid range expansion and lack of acquired immunity. In addition, rapid habitat loss and urbanization in Florida may increase the probability of disease transmission between wild and domestic canids. We suggest preventative measures that may lower the risk of parasitic infection and promote co-existence with coyotes in urban landscapes.

Cognitive Dysfunction in Stroke Survivors: A Community-Based Prospective Study from Kolkata, India

The frequency of cognitive dysfunction among community stroke survivors (SS) is not known in India. This prospective study investigated the prevalence of poststroke mild cognitive impairment (psMCI) and poststroke dementia (psDem), the annual progression rate to dementia, and pertinent risk factors in a sample population of SS in Kolkata, India between September 2006 and July 2010. From a community-based stroke registry, 281 SS were assessed at baseline year after excluding attrition due to various causes. Validated cognitive tools were applied by trained field workers under supervision of a neuropsychologist. The assessment was repeated annually for consecutive three years. The mean follow-up period was 1.89 years (range, 1-3 years). The period prevalence rate of psDem was 13.88% (95% confidence interval [CI], 9.91%-18.90%) at baseline, and the average annual progression rate to dementia was 3.53% (95% CI, 2.09%-5.58%). Compared with subjects without dementia, those with psDem were significantly older at first-ever stroke and more likely to have cortical atrophy. The period prevalence rate of psMCI was 6.05% (95% CI, 1.45%-13.64%) at baseline, and 10.6% (95% CI, 4.57%-20.88%) of these subjects converted to psDem annually. Survival analysis of psDem patients showed a greater risk of death in psDem SS as compared to nondemented SS (hazard ratio, 2.65; 95% CI, 1.72-6.15). Our data suggest that the overall prevalence of psDem is higher than that of psMCI, possibly related to nonexclusion of prestroke dementia, but that the average annual progression rate of SS to psDem and that of psMCI to psDem are comparable. Older age at first-ever stroke and cortical atrophy are associated with increased risk for dementia among SS.