Whistle discrimination and categorization by the Atlantic bottlenose dolphin (Tursiops truncatus): a review of the signature whistle framework and a perceptual test

Dolphin whistles vary by frequency contour, changes in frequency over time. Individual dolphins may broadcast their identities via uniquely contoured whistles, "signature whistles." A recent debate concerning categorization of these whistles has highlighted the on-going need for perceptual studies of whistles by dolphins. This article reviews research on dolphin whistles as well as presenting a study in which a captive, female, adult bottlenose dolphin performed a conditional matching task in which whistles produced by six wild dolphins in Sarasota Bay were each paired with surrogate producers, specific objects/places. The dolphin subject also categorized unfamiliar exemplars produced by the whistlers represented by the original stimuli. The dolphin successfully discriminated among the group of whistles, associated them with surrogate producers, grouped new exemplars of the same dolphin's whistle together when the contour was intact, and discriminated among same-contour whistles produced by the same dolphin. Whistle sequences that included partial contours were not categorized with the original whistlers. Categorization appeared to be based on contour rather than specific acoustic parameters or voice cues. These findings are consistent with the perceptual tenets associated with the signature whistle framework which suggests that dolphins use individualized whistle contours for identification of known conspecifics.     

A comparison of pectoral fin contact behaviour for three distinct dolphin populations

Tactile exchanges involving the pectoral fin have been documented in a variety of dolphin species. Several functions (e.g., social, hygienic) have been offered as possible explanations for when and why dolphins exchange pectoral fin contacts. In this study, we compared pectoral fin contact between dolphin dyads from three distinct dolphin populations: two groups of wild dolphins; Atlantic spotted dolphins (Stenella frontalis) from The Bahamas and Indo-Pacific bottlenose dolphins (Tursiops aduncus) from around Mikura Island, Japan; and one group of captive bottlenose dolphins (Tursiops truncatus) residing at the Roatan Institute for Marine Sciences, Anthony's Key Resort. A number of similarities were observed between the captive and wild groups, including; rates of pectoral fin contact, which dolphin initiated contact, posture preference, and same-sex rubbing partner preference. Unlike their wild counterparts, however, dolphins in the captive study group engaged in petting and rubbing at equal rates, females were more likely to contact males, males assumed the various rubbing roles more frequently than females, and calves and juveniles were more likely to be involved in pectoral fin contact exchanges. These results suggest that some aspects of pectoral fin contact behaviour might be common to many dolphin species, whereas other aspects could be species specific, or could be the result of differing environmental and social conditions.     

Quantitative tools for comparing animal communication systems: information theory applied to bottlenose dolphin whistle repertoires

Comparative analysis of nonhuman animal communication systems and their complexity, particularly in comparison to human language, has been generally hampered by both a lack of sufficiently extensive data sets and appropriate analytic tools. Information theory measures provide an important quantitative tool for examining and comparing communication systems across species. In this paper we use the original application of information theory, that of statistical examination of a communication system's structure and organization. As an example of the utility of information theory to the analysis of animal communication systems, we applied a series of information theory statistics to a statistically categorized set of bottlenose dolphin Tursiops truncatus, whistle vocalizations. First, we use the first-order entropic relation in a Zipf-type diagram (Zipf 1949 Human Behavior and the Principle of Least Effort) to illustrate the application of temporal statistics as comparative indicators of repertoire complexity, and as possible predictive indicators of acquisition/learning in animal vocal repertoires. Second, we illustrate the need for more extensive temporal data sets when examining the higher entropic orders, indicative of higher levels of internal informational structure, of such vocalizations, which could begin to allow the statistical reconstruction of repertoire organization. Third, we propose using 'communication capacity' as a measure of the degree of temporal structure and complexity of statistical correlation, represented by the values of entropic order, as an objective tool for interspecies comparison of communication complexity. In doing so, we introduce a new comparative measure, the slope of Shannon entropies, and illustrate how it potentially can be used to compare the organizational complexity of vocal repertoires across a diversity of species. Finally, we illustrate the nature and predictive application of these higher-order entropies using a preliminary sample of dolphin whistle vocalizations. The purpose of this preliminary report is to re-examine the original application of information theory to the field of animal communication, illustrate its potential utility as a comparative tool for examining the internal informational structure of animal vocal repertoires and their development, and discuss its relationship to behavioural ecology and evolutionary theory. Copyright 1999 The Association for the Study of Animal Behaviour.     

Automated classification of bird and amphibian calls using machine learning: A comparison of methods

We compared the ability of three machine learning algorithms (linear discriminant analysis, decision tree, and support vector machines) to automate the classification of calls of nine frogs and three bird species. In addition, we tested two ways of characterizing each call to train/test the system. Calls were characterized with four standard call variables (minimum and maximum frequencies, call duration and maximum power) or eleven variables that included three standard call variables (minimum and maximum frequencies, call duration) and a coarse representation of call structure (frequency of maximum power in eight segments of the call). A total of 10,061 isolated calls were used to train/test the system. The average true positive rates for the three methods were: 94.95% for support vector machine (0.94% average false positive rate), 89.20% for decision tree (1.25% average false positive rate) and 71.45% for linear discriminant analysis (1.98% average false positive rate). There was no statistical difference in classification accuracy based on 4 or 11 call variables, but this efficient data reduction technique in conjunction with the high classification accuracy of the SVM is a promising combination for automated species identification by sound. By combining automated digital recording systems with our automated classification technique, we can greatly increase the temporal and spatial coverage of biodiversity data collection.     

Sampling Atriplex bladders: a comparison of methods

Abstract. A major objection to the convenient sampling of Chenopodiacean bladders by brushing in aqueous solutions is the suspected leakage of ions from the leaf blade. To overcome this problem, a new method of sampling bladder hair is described, which also achieves rapid and quantitative separation and, at the same time, yields bladder samples of undoubtedly high purity. Leakage is stopped by the aid of liquid nitrogen. Comparison of this method to removal of bladders by brushing in aquenous solutions effectively confirm the validity of the customary method beyond dispute.     

Disrupting Escherichia coli: a comparison of methods

The often-encountered problem of disrupting bacteria for the purpose of extracting soluble protein has generated various methods. Many require specialized equipment. Very often, especially during preliminary studies, investigators need a simple, fast, and inexpensive method for cell disruption that preserves biological activity. This paper compares some simple and inexpensive methods for cell disruption, such as bead-vortexing, freezing-thawing, French pressing, and sonication. It also provides some tips to increase protein yield and preserve biological activity. If performed under optimal conditions, bead-vortexing gives protein yields that are comparable to French pressing and sonication. It also preserves the activities of labile enzymes and releases periplasmic enzymes. Vortexing with glass beads appears to be the simplest method for cell disruption.     

Protein determination: a comparison of several methods

A comparison of four methods for the determination of total proteins is presented from the following points of view: - sensitivity; - specificity; - amount of work, chemicals, time and equipment needed for the performance of the determination. The following tests have been examined; Tombs' (absorbancy at 210 nm); Waddell's (difference in absorbancy between 215 and 225 nm); Warburg's (absorbancy at 280 nm); Lowry's (absorbancy at 500 nm after the reaction with phenol reagent). The authors recommend Tombs' method for its outstanding sensitivity, specificity and simplicity as the best of the four.     

Gonorrhea modeling: a comparison of control methods

A population dynamics model for a heterogeneous population is used to compare the effectiveness of six prevention methods for gonorrhea involving population screening and contact tracing of selected groups. The population is subdivided according to sex, sexual activity, and symptomatic or asymptomatic infection. For this model contact tracing of certain groups is more effective than general population screening.     

Flights of fear: a mechanical wing whistle sounds the alarm in a flocking bird

Animals often form groups to increase collective vigilance and allow early detection of predators, but this benefit of sociality relies on rapid transfer of information. Among birds, alarm calls are not present in all species, while other proposed mechanisms of information transfer are inefficient. We tested whether wing sounds can encode reliable information on danger. Individuals taking off in alarm fly more quickly or ascend more steeply, so may produce different sounds in alarmed than in routine flight, which then act as reliable cues of alarm, or honest ‘index’ signals in which a signal''s meaning is associated with its method of production. We show that crested pigeons, Ocyphaps lophotes, which have modified flight feathers, produce distinct wing ‘whistles’ in alarmed flight, and that individuals take off in alarm only after playback of alarmed whistles. Furthermore, amplitude-manipulated playbacks showed that response depends on whistle structure, such as tempo, not simply amplitude. We believe this is the first demonstration that flight noise can send information about alarm, and suggest that take-off noise could provide a cue of alarm in many flocking species, with feather modification evolving specifically to signal alarm in some. Similar reliable cues or index signals could occur in other animals.     

Virginity in haplodiploid populations: a comparison of estimation methods

1. The presence of ovipositing virgin females, who are able to produce only male offspring, may have several important consequences for the evolution of reproductive strategies.
2. The prevalence of virginity for five fig wasp species was estimated using three different methods: (1) the proportion of females developing in fruit that contained no conspecific males; (2) dissection of females caught on sticky traps; and (3) dissection of females that had emerged from their galled flowers into the fruit cavity.
3. The estimates obtained by method 1 (females developing in single sex broods) were lower than those obtained by the other methods.
4. Across species, the estimates obtained by method 1 (females developing in single sex broods) were significantly correlated with those obtained by method 2 (sticky trap caught females).
5. Mating with sperm-depleted males is unlikely to occur in these species.     

Assessing parental effort in a Neotropical parrot: a comparison of methods

Various behavioural or body condition indices are frequently used to measure avian parental effort. To determine which of the indices provide meaningful short-term assessments of parental effort, we used four of them concurrently on breeding green-rumped parrotlets, Forpus passerinus, and then compared the results with field metabolic rate (FMR) measurements obtained using the doubly labelled water (DLW) technique. Nest provisioning rate correlated significantly with FMR among both males and females, while proportion of time away from the nest was correlated with FMR in females but not in males. Adult mass change during the breeding cycle and feeding latency were both unrelated to FMR. Our results suggest researchers should be cautious in assuming that any indirect method of quantifying parental effort is reliable for a particular species. We also tested whether the relatively invasive DLW protocol causes potentially confounding changes in study animals' subsequent behaviour. Paired observations before and after DLW injection showed no measurable changes in parrotlet nest attendance patterns or nest provisioning rate. Copyright 1999 The Association for the Study of Animal Behaviour.     

Carbon-specific phytoplankton growth rates: a comparison of methods

Measurements of biomass and growth rate of two axenic algalcultures were carried out using three different methodologicalapproaches: the specific ¹⁴C-labelling of chlorophyll a, [³H]adenineincorporation into DNA and net organic carbon assimilation.Time-course experiments revealed that the specific activitiesof chlorophyll a were significantly higher than the specificactivity of total algal carbon in six of seven experiments.When the specific activity of chlorophyll a is used to calculatethe carbon biomass and growth rate, the carbon biomass of thealgae will thus be underestimated and the specific growth ratewill be too high. Determination of growth rates from incorporationof [³H]adenine gave lower values than those obtained from netorganic carbon assimilation and from ¹⁴C incorporation intochlorophyll a. Problems with adenine saturation are suggested.When [³H]adenine is used to measure growth rates in dense algalcultures, additions of >1 µM [³H]adenine are oftenrequired to maximally label the extracellular and intracellularadenine pools and hence DNA.     

Prediction error estimation: a comparison of resampling methods

MOTIVATION: In genomic studies, thousands of features are collected on relatively few samples. One of the goals of these studies is to build classifiers to predict the outcome of future observations. There are three inherent steps to this process: feature selection, model selection and prediction assessment. With a focus on prediction assessment, we compare several methods for estimating the 'true' prediction error of a prediction model in the presence of feature selection. RESULTS: For small studies where features are selected from thousands of candidates, the resubstitution and simple split-sample estimates are seriously biased. In these small samples, leave-one-out cross-validation (LOOCV), 10-fold cross-validation (CV) and the .632+ bootstrap have the smallest bias for diagonal discriminant analysis, nearest neighbor and classification trees. LOOCV and 10-fold CV have the smallest bias for linear discriminant analysis. Additionally, LOOCV, 5- and 10-fold CV, and the .632+ bootstrap have the lowest mean square error. The .632+ bootstrap is quite biased in small sample sizes with strong signal-to-noise ratios. Differences in performance among resampling methods are reduced as the number of specimens available increase. SUPPLEMENTARY INFORMATION: A complete compilation of results and R code for simulations and analyses are available in Molinaro et al. (2005) (http://linus.nci.nih.gov/brb/TechReport.htm).     

Restoration of dredged canals in wetlands: a comparison of methods

A comparison of two methods for restoring dredged canals to wetlands was examined at the Jean Lafitte National Historical Park and Preserve’s Barataria Preserve Unit near New Orleans, LA. Both northern and southern canals had the remnant dredged spoil material returned to the canal, but the southern canal had additional sediment pumped in from a nearby lake. The water depth in the southern canal shallowed significantly from 1.2 to 0.4 m following backfilling and sediment addition, while the depth of the northern canal (which received no additional sediment) remained unchanged following backfilling. Neither site had complete soil restoration, but the former spoil areas of the northern canal showed greater restoration than the southern canal. The vegetation on the former spoil areas of the northern canal closely resembled that of the reference marsh, while the former spoil areas of the southern canal had species indicative of spoil banks and other elevated areas. After 3 years wetland vegetation was established on approximately 65% of the former spoil areas at both sites and 20–25% of the open water areas. Sediment addition to the southern canal raised costs by a factor of eight times compared to that of the northern canal. The results of this study document the restoration potential of both methods, but also show that backfilling without supplemental sediment additions can restore abandoned canals at a fraction of the cost of other methods.     

Reproductive behaviour in survival: a comparison between wild and domestic sheep

The recorded behaviour of wild species of Ovis at the beginning of the breeding season supports the proposal that mating is synchronized by the 'ram effect', a phenomenon already described in domestic sheep. Animals separate into exclusive male flocks, and flocks of females and young animals for most of the year. They reunite just before the rutting season. At lambing there appear to be behavioural mechanisms that ensure that ewes lamb in close proximity to one another. It is hypothesized that these behavioural characteristics of wild sheep help protect the newborn and that much of the reproductive and maternal behaviour of domestic sheep may be traced to comparable behaviour in wild species.