International Journal of Primatology

, Volume 37, Issue 2, pp 281–295 | Cite as

Deadwood Structural Properties May Influence Aye-Aye (Daubentonia madagascariensis) Extractive Foraging Behavior

  • Katharine E. T. Thompson
  • Richard J. Bankoff
  • Edward E. LouisJr.
  • George H. Perry
Article

Abstract

The identification of critical, limited natural resources for different primate species is important for advancing our understanding of behavioral ecology and toward future conservation efforts. The aye-aye (Daubentonia madagascariensis) is an Endangered nocturnal lemur with adaptations for accessing structurally defended foods: continuously growing incisors; an elongated, flexible middle finger; and a specialized auditory system. In some seasons, ca. 90% of the aye-aye’s diet consists of two structurally defended resources: 1) the larvae of wood boring insects, extracted after the aye-aye gnaws through decomposing bark (deadwood), and 2) the seeds of Canarium trees. Aye-ayes have very large individual home ranges relative to most other lemurs, possibly owing to limited resource availability. Identification of limiting dietary factor(s) is critical for our understanding of aye-aye behavioral ecology and future conservation efforts. To investigate whether aye-ayes equally access all deadwood resources within their range, we surveyed two 100 × 100 m forest plots within the territories of two aye-ayes at Sangasanga, Kianjavato, Madagascar. Only 2 of 150 deadwood specimens within the plots (1.3%) appeared to have been accessed by the aye-ayes. To test whether any external or internal deadwood properties explain aye-aye foraging preferences we recorded species, height and diameter, and quantified the internal tree density using a 3D acoustic tomograph for each foraged and nonforaged deadwood resource within the plots, plus 13 specimens (5 foraged and 8 nonforaged) outside the plots. We did not detect any statistically significant preferences for species, diameter, or height. However, results from the acoustic analysis tentatively indicated that aye-ayes are more likely to forage in trees with greater internal (≥6 cm from the bark) densities. This interior region may function as a sounding board in the tap-foraging process to help aye-ayes accurately identify potential grub-containing cavities in the outer 1–4 cm of deadwood.

Keywords

Food resource limitation Home range Percussive foraging Structurally defended food resources 

Introduction

Resource availability can be a major predictor of home range size for primate species (Di Bitetti 2001). For example, among lemurs in Madagascar, Milne-Edwards’ sifakas (Propithecus edwardsi) have relatively larger home ranges in environmentally disturbed, patchy habitats than conspecifics in primary habitats (Gerber et al. 2012). Similarly, ring-tailed lemurs (Lemur catta) in Beza Mahafaly expand their home ranges during the dry season, likely to accommodate seasonal resource scarcity (Sussman 1991).

In addition to expanded home range sizes, low forest productivity and patchy resource distribution may lead to decreased population densities (Hanya and Chapman 2012). Therefore, instances of solitary foraging behavior over large home ranges may be a consequence of food limitation. For example, it is hypothesized that the large, solitary home ranges of Bornean orangutans (Pongo pygmaeus) are an adaptive response to satisfying large nutrient requirements in the face of scarce food resources (Galdikas 1988; Mackinnon 1974; Singleton and van Schaik 2001).

We here investigated potential food resource limitations in the wild habitat of a solitary foraging primate with extensive home range sizes, the aye-aye (Daubentonia madagascariensis). The aye-aye is an Endangered nocturnal lemur endemic to Madagascar with an adult body mass of ca. 2.5 kg and the largest relative brain size of any extant strepsirrhine primate (Kaufman et al. 2005; Schwitzer et al. 2013; Smith and Jungers 1997; Sterling 1993). Aye-ayes are able to access structurally defended foods with the aid of number of morphological adaptations, including continuously growing incisors; an elongated, thin, and highly flexible middle finger; and a specialized auditory system (Cartmill 1974; Simons 1995; Sterling 1993).

The longest aye-aye behavioral study to date was conducted from 1989 to 1991 on Nosy Mangabe, an island to which nine aye-ayes were introduced in the 1960s (Sterling 1993, 1994). At this site, up to ca. 90% of the aye-aye's diet during hot-dry seasons consisted of two structurally defended resources: the larvae of wood-boring insects (extracted from decomposing bark and wood, i.e., deadwood) and the seeds of Canarium trees (Sterling 1993, 1994). On Nosy Mangabe and at other sites, smaller components of the aye-aye diet include tree sap, nectar, fruits, and adult insects (Iwano and Iwakawa 1988; Pollock et al. 1985; Sterling 1993, 1994).

Aye-ayes extract insect larvae from a variety of locations, including trunks and branches of dead trees, fallen deadwood, dead branches on living trees, bamboo, and —very rarely— living trees (Sefczek et al. 2012; Sterling 1994). The middle phalanx is used to tap on the bark of decaying trees to facilitate the detection of tree cavities that may contain grubs, in a process referred to as percussive foraging (Erickson 1994). Aye-ayes rely on the auditory interpretation of signals that are produced by rapidly tapping the outer surface of the tree to identify larval mines beneath the bark (Coleman and Ross 2004; Erickson 1994; Kaufman et al. 2005; Ramsier and Dominy 2012). The continuously growing incisors are then used to gouge holes in the trees, from which the insect larvae are skewered with the flexible middle digit (Erickson 1994; Sterling 1994).

Although insect larvae are often envisioned as the most important aye-aye food, researchers have alternatively suggested the nutrient-rich seeds of Canarium as a primary dietary resource (Iwano and Iwakawa 1988; Sterling 1993). Important Canarium species for aye-ayes include C. boivinii and C. madagascariensis (Iwano and Iwakawa 1988; Pollock et al. 1985; Simons 1995; Sterling 1993, 1994). To access the nut-like seeds of Canarium, aye-ayes use their superior incisors to chew through the mid-endocarp and their inferior incisors to pierce the endocarp before scraping out the interior with their middle finger (Sterling 1994).

The aye-aye has the largest geographical species distribution of any extant lemur, with populations observed in both the tropical rainforests along the east coast and the relatively dry forests of the west coast and northernmost regions of Madagascar (Sterling 2003). Aye-ayes also have among the largest individual home range (the core area of land within which an animal forages) sizes of any lemur: 120–215 ha for males and 30–40 ha for females on Nosy Mangabe (Sterling 1993). These home range sizes are extensive given that aye-ayes are solitary foragers with a ca. 2.5-kg body size (Smith and Jungers 1997; Sterling 1993). Although aye-ayes have smaller body sizes than orangutans, like orangutans they have relatively large, minimally overlapping (at least for aye-aye females) home ranges and solitary foraging behaviors that could reflect dependence on one or more limited food resources.

Studying potential food resource limitation is important for our understanding of aye-aye behavioral ecology and for future conservation efforts. With deforestation and the further loss of deadwood resources or live Canarium trees (the densities of which may vary across habitats), aye-ayes may be forced to range farther to satisfy dietary needs to avoid local extirpation (Farris et al. 2011; Sefczek et al.2012). In this study, we performed an initial investigation into potentially limiting factors of the aye-aye diet, focusing especially on deadwood. Specifically, to test whether deadwood in general is a limited resource we first assessed whether aye-ayes exploited all or a smaller portion of the available deadwood within their home ranges in order. Second, we investigated whether particular external or internal properties of deadwood or the proximity to living Canarium trees may affect aye-aye resource selection.

Methods

Study Site

We collected data from June to August 2013 in the Sangasanga forest near the village of Kianjavato. Kianjavato (Fig. 1a) is located in the Vatovavy–Fitovinany Region of the Fianarantsoa Province in southeastern Madagascar (latitude: –21.37 and longitude: 47.87) (Schwitzer et al. 2013). The elevation at Sangasanga ranges from 52 to 571 m (Manjaribe et al. 2013). At its lowest elevation the forest borders the village of Kianjavato and is thus frequently used by local people. This area is less densely forested than areas of higher elevation as it is intersected by informal pathways, bamboo-lined streams, and coffee plantations managed by Foibe Fihofanana momba ny Fambolena (FOFIFA; The Plantation Management Headquarters) (Schwitzer et al. 2013). Above this area (at higher elevation) there is a denser band of forest composed of a thick understory that experiences less human disturbance. Toward the area’s highest peaks the forest consists predominantly of the traveler’s palm, Ravenala madagascariensis.
Fig. 1

Study plots and site in Kianjavato, Madagascar (June–August 2013). (A) Location of Kianjavato, Madagascar. (B) Spatial distribution of aye-aye-foraged and nonforaged trees. GPS locations of all deadwood standing or fallen tree resources (including stumps and logs) that we analyzed in this study within the Sangasanga forest in Kianjavato. Red denotes the presence of feeding traces on a specimen, while blue indicates the tree was not foraged upon. Circles and triangles indicate specimens found with 100 × 100 m plots; squares indicate supplementary specimens from nearby the plots included to increase sample size (five additional foraged trees and eight nonforaged trees). For the figure, a small number of points that we recorded improperly on the GPS unit were positioned manually based on field notes; these data are used for visualization purposes only and do not affect any analyses. Plots were 0.1528 km apart at the midpoint, and points outside the plot were on average 0.0919 km from the midpoint of plot 2.

The Madagascar Biodiversity Partnership at the Kianjavato Ahmanson Field Station actively monitors and collects behavioral data on GPS- and radio-collared aye-ayes at Sangasanga. Our specific study plots (Fig. 1b) were located within the active home ranges of two collared adult aye-aye individuals (one adult male and one adult female; the female had a single offspring during the time of the study), in a zone of humid secondary lowland rainforest ca. 0.69 km from the village, the periphery of which is adjacent to the forest boundary. The collared aye-ayes range across all of the aforementioned forest regions (Manjaribe et al. 2013; Solofondranohatra 2014). The well-documented ranging patterns and foraging habits of aye-ayes within this area, along with Sangasanga’s topographic, botanical, and level of anthropogenic disturbance variation made it an ideal site for this study.

Inventory of Tree Distribution

Our goal was to perform an initial test of potentially limiting food resources in the aye-aye home range, including of deadwood in general, particular external or internal properties of deadwood, and the proximity of deadwood to live Canarium trees. Densities of Canarium are largely unquantified in southeastern Madagascar rainforests, with the limited available data suggesting a patchy pattern of distribution (Farris et al. 2011; Sefczek et al. 2012). Therefore, we chose to survey two relatively large plots (100 × 100 m each) rather than a greater number of smaller plots or transects, hoping that the plots would each encompass multiple specimens of Canarium.

We established one 100 × 100 m plot in the low-elevation, disturbed forest zone and a second 100 × 100 m plot in the medium to high-elevation, dense forest zone. We placed plots within areas of frequent use in the home ranges of the two radio-collared individuals, and far from any of the ranges’ boundaries. We recorded the diameter (specifically diameter at breast height [DBH]), height, vernacular name, and GPS location (using a Garmin GPSmap 62) of each deadwood resource and living Canarium tree specimen. We measured DBH using two observers of identical stature (160 cm with DBH measured at ca. 119 cm). At breast height, the researchers took measurements using a DBH forestry measuring tape (which displays both the circumference and diameter). In the event that deadwood resources were <119 cm in height, diameter was taken at the highest available point.

We measured tree heights directly using a meter tape where the full length of the specimen was accessible or could be safely climbed, and visually estimated when part of the specimen was inaccessible or beyond the reach of equipment. We identified all deadwood and Canarium specimens according to Malagasy vernacular names by bark and appearance. We later translated vernacular names to scientific names whenever possible. We included in this analysis only the deadwood specimens whose Malagasy vernacular names could be translated into scientific names. Of the translatable specimens, 23 could be identified only to the genus level; we therefore performed the analysis at this taxonomic level.

We identified all traces of aye-aye foraging for larvae, typically in deadwood. We also searched for traces on live trees, which are very occasional targets of aye-aye foraging (Sefczek et al.2012). Signs of foraging events included incisor indentations with funnel-shaped holes at the center of gnawed areas and bark shavings peeled away from the point of insertion; these traces served as proxies for direct foraging observations at a given tree (Duckworth 1993; Erickson 1994; Sefczek et al. 2012; Sterling 1993). We surveyed each deadwood resource extensively and on all sides, in daylight, by multiple members of the research team, including with binoculars. We are confident that we were able to identify the vast majority of aye-aye traces within the plots. Still, it is possible that we missed very small traces that were located higher in the canopy. Any such sampling error is expected to be small and is not anticipated to affect the overall conclusions of the study. Within the two 100 × 100 m plots, we identified a total of two foraged and 148 nonforaged deadwood resources. To increase our sample size of foraged deadwood for our external and internal tree property analyses, we additionally identified and examined a further five deadwood resources with aye-aye foraging traces and eight nonforaged deadwood resources that were outside the two fully surveyed 100 × 100 m plots but that were still located within the known ranges of the collared aye-aye individuals. These additional traces represented the most proximate instances of foraging beyond the boundaries of the plot but well within the aye-aye’s home range, plus nearby nonforaged specimens. The expanded sample size thus consisted of a total of 163 deadwood tree resources (foraged N = 7 and nonforaged N = 156).

Deadwood Internal Structural Property Measurement

Given the mechanics of the percussive “tap” foraging method of aye-aye larval extraction, variation in deadwood structural qualities may influence foraging ability or success. To estimate the internal density of foraged vs. nonforaged deadwood resources, we used a Fakopp Arborsonic 3D Acoustic Tomograph that estimates the velocity at which sound is conducted through the tree. This forestry machine distributes and receives sound waves across 10 sensors placed evenly around the circumference of a tree. Each sensor has thin, elongated prongs with precise sound sensors at the tip; these prongs are hammered into the bark of the tree. Each sensor is then tapped a minimum of three times, which propagates sound across the tree. The sensors are connected to a laptop, where the Fakopp Arborsonic software (with custom modifications for our study, Version 3) records the length of time taken for the sound to be received by each sensor. Sound travels at different speeds depending on the density of the substrate. It traverses more quickly through tightly packed (dense) molecules and more slowly across loosely packed molecules. The Arborsonic program renders two- and three-dimensional computer models of internal tree density based on the velocity of the sound traveling between the sensors. Therefore, scanning the deadwood specimens allowed us to quantify internal structural property variation that may influence aye-aye tap-foraging behavior.

For each deadwood tree resource, we performed five scans between 80- and 120-cm tree heights at ca. 5-cm intervals for nonforaged and ca. 2-cm intervals for foraged specimens. For shorter specimens, deeply decayed spots, or other tree malformations, we scanned as close as possible to the breast height of the tree. For foraged specimens, we took the five scans around the sites of aye-aye traces. The external bark of 2 of the 7 foraged and 109 of the 156 nonforaged deadwood tree resources were too decayed to hold the Acoustic Tomograph prongs tightly, precluding the ability to obtain accurate sound velocity data. For a minority (N = 7) of the nonforaged specimens included in the sample, it was not possible to collect a full five scans (Electronic Supplementary Material [ESM] Table SI). Our final sample sizes for the internal structural property analysis were thus 25 scans from 5 foraged trees and 224 scans from 47 nonforaged trees.

For each scan, we used the Arborsonic software (http://www.fakopp.com/site/download) to generate a 101 × 101 velocity matrix and corresponding XY coordinates (in cm) for each velocity measurement. From these data, we used weighted averages of the values nearest to points at discrete (1-cm interval) tree depths from the outer surface of the tree along four X- and Y-axes toward the origin of the tree. That is, we produced four values per scan for each 1-cm interval from depths of 1–10 cm, with the number of intervals for each scan dependent on the radius of the tree at the scanning location. With these data we then compared velocity estimates at discrete tree depths between foraged and nonforaged specimens. The density values from a given region of one deadwood specimen are likely related to the density values of a different region within that same tree. That is, for the aforementioned analysis we recorded four velocity values per cm interval from each of five scans per tree. Therefore, we in addition calculated a single average velocity value for each cm interval per tree and repeated the analysis. The means and t-test parameters for the external and internal variables of deadwood analyses are provided in Table I. The original and processed Arborsonic files for each scan and the R scripts used for these analyses have been deposited in the Dryad Digital Repository at http://dx.doi.org/10.5061/dryad.d4p7n.
Table I

Means and test parameters for foraged vs. nonforaged deadwood comparisons

Variable

Nonforaged trees

Foraged trees

Nonforaged mean (±SE) [±SD]

Foraged mean (±SE) [±SD]

t-test results

95% CI of the differencea

N trees

N values tested

N trees

N values tested

t

df

P

Lower

Upper

Height

153

153

7

7

5.5 (±0.6) [±7.3]

6.9 (±2.2) [±5.8]

–0.6

7

0.57

–6.7

4

Diameter (of all specimens)

150

150

7

7

23 (±1.3) [±15.5]

18.9 (±4.1) [±10.7]

1.0

8

0.36

–5.8

14.1

Diameter measured specifically at breast height (ca. 119 cm)

94

94

7

7

26.7 (±1.9) [±17.6]

18.9 (±4.1) [±4.1]

1.8

9

0.11

–2.3

17.9

Velocity

 at 1 cm (all scan values)

27

480

4

80

1544.4 (±28.1) [±614.3]

1217.4 (±58.5) [±522.6]

–5.1

119

1.65E–06

–455.4

–198.7

 at 2 cm (all scan values)

27

480

4

80

1598.8 (±28.6) [±625.5]

1366.3 (±74.5) [±665.9]

–3

104

0.004

–390.7

–74.5

 at 3 cm (all scan values)

41

672

5

100

1491.9 (±23.2) [±600.3]

1326.1 (±56.8) [±567.2]

–2.8

135

0.01

–287

–44.7

 at 4 cm (all scan values)

44

812

3

60

1372 (±21.4) [±607.9]

1370.4 (±74.3) [±574.8]

–0.1

70

0.99

–155.7

152.4

 at 5 cm (all scan values)

43

788

3

60

1298 (±22.3) [±623.6]

1414.4 (±70.5) [±545.8]

1.6

72

0.12

–30.9

263.8

 at 6 cm (all scan values)

41

764

3

60

1223.4 (±22.4) [±618.5]

1433.3 (±66.9) [±518.2]

3.0

73

0.004

69.3

350.5

 at 7 cm (all scan values)

38

708

3

60

1116 (±19.7) [±522.5]

1433.3 (±64.7) [±500.5]

4.7

71

1.26E–05

182.7

452

 at 8 cm (all scan values)

35

624

2

36

1026.1 (±18.9) [±470]

1440.6 (±98.2) [±589.3]

4.2

38

0.0002

212.1

617.1

 at 1 cm (measured at 80–120 cm)

11

112

2

44

1547.9 (±65.5) [±692.7]

1498.3 (±85.6) [±567.4]

–0.5

96

0.65

–263.5

164.2

 at 2 cm (measured at 80–120 cm)

11

112

2

44

1594 (±67.7) [±715.9]

1817.6 (±105.7) [±700.8]

1.8

81

0.08

–26.1

473.3

 at 3 cm (measured at 80–120 cm)

14

144

3

64

1500.9 (±56.6) [±678.8]

1598.2 (±76.7) [±612.9]

1.1

134

0.31

–91.1

285.7

 at 4 cm (measured at 80–120 cm)

16

188

2

40

1355.5 (±49.3) [±675.3]

1530.1 (±94.4) [±597]

1.7

63

0.11

–38.3

387.4

 at 5 cm (measured at 80–120 cm)

14

188

2

40

1213.6 (±51) [±699]

1567.3 (±89.2) [±563.9]

3.5

68

9.91E–04

148.8

558.7

 at 6 cm (measured at 80–120 cm)

13

204

2

40

1128.9 (±48.5) [±692.3]

1575.7 (±84.6) [±535]

4.6

68

2.03E–05

252.4

641.5

 at 7 cm (measured at 80–120 cm)

12

184

2

40

942.6 (±34.2) [±463]

1567.5 (±81.8) [±517.3]

7.1

54

3.58E–09

447.2

802.6

 at 8 cm (measured at 80–120 cm)

11

168

1

20

835.5 (±25.8) [±333.7]

1581.6 (±161.3) [±721]

4.6

20

1.86E–04

405.6

1086.8

 at 1 cm (one average value per deadwood specimen)

27

27

4

4

1515.7 (±92.3) [±479.3]

1217.4 (±174.5) [±349]

–1.6

5

0.2

–809.8

213.2

 at 2 cm (one average value per deadwood specimen)

27

27

4

4

1565.4 (±100.3) [±521.2]

1366.3 (±239.5) [±478.9]

–0.8

5

0.49

–911

512.7

 at 3 cm (one average value per deadwood specimen)

41

41

5

5

1426 (±82.7) [±529.5]

1326.1 (±167.1) [±373.6]

–0.6

7

0.62

–553

353.4

 at 4 cm (one average value per deadwood specimen)

44

44

3

3

1358.3 (±82.7) [±548.1]

1370.4 (±189.2) [±327.7]

0.1

3

0.96

–667.7

691.9

 at 5 cm (one average value per deadwood specimen)

43

43

3

3

1293.4 (±87.4) [±573]

1414.4 (±164.3) [±284.5]

0.7

4

0.56

–443.4

685.4

 at 6 cm (one average value per deadwood specimen)

41

41

3

3

1241.4 (±91.4) [±584.7]

1433.3 (±146.2) [±253.2]

–1.2

4

0.34

–678.6

294.7

 at 7 cm (one average value per deadwood specimen)

38

38

3

3

1141.8 (±81.5) [±505.7]

1433.3 (±135.5) [±234.7]

–1.9

4

0.15

–745.7

162.8

 at 1 cm (measured at 80–120 cm, one average value per deadwood specimen)

11

11

2

2

1471.1 (±165) [±547.1]

1506.8 (±93.6) [±132.3]

0.2

9

0.86

–396.5

467.8

 at 2 cm (measured at 80–120 cm, one average value per deadwood specimen)

11

11

2

2

1517.9 (±186.2) [±617.5]

1808.4 (±101.2) [±143.2]

1.4

9

0.21

–189.2

770.1

 at 3 cm (measured at 80–120 cm, one average value per deadwood specimen)

14

14

3

3

1426.8 (±159.2) [±595.7]

1582.9 (±194.2) [±336.3]

0.7

6

0.56

–480.7

793.1

 at 4 cm (measured at 80–120 cm, one average value per deadwood specimen)

16

16

2

2

1334.3 (±152.1) [±608.3]

1530.1 (±175.8) [±248.6]

0.9

3

0.47

–551.5

943

 at 5 cm (measured at 80–120 cm, one average value per deadwood specimen)

14

14

2

2

1207 (±177.4) [±663.6]

1567.3 (±103.8) [±146.8]

1.8

10

0.12

–102.5

823.1

 at 6 cm (measured at 80–120 cm, one average value per deadwood specimen)

13

13

2

2

1140.2 (±182.4) [±657.5]

1575.7 (±56.7) [±80.1]

–2.3

13

0.05

–848.1

–23

 at 7 cm (measured at 80–120 cm, one average value per deadwood specimen)

12

12

2

2

999.8 (±137.8) [±477.1]

1567.5 (±31.8) [±45]

–4.1

12

0.002

–876.1

–259.3

a95% CI = the confidence interval for the difference between the nonforaged means and foraged means, i.e., nonforaged – foraged, such that a negative value represents a greater foraged than nonforaged value.

Results

Deadwood Genus and External Properties

Within the two 100 × 100 m forest plots, only 2 of 150 total observed deadwood tree specimens (1.3%) had been accessed by the aye-ayes for insect larvae on the basis of visible signs of past extractive foraging (Fig. 1b). This result suggests that deadwood alone, i.e., without the consideration of any other variables, is not likely a limited resource for aye-ayes.

For all subsequent analyses we established a combined database of total foraged (N = 7) and total nonforaged (N = 156) deadwood resources (N = 163 specimens total) within and nearby outside the study plots (Fig. 1b). We observed that the deadwood foraging events were distributed across six different tree genera; thus there was no statistically significant preference for particular taxa relative to the distribution of nonforaged deadwood resources (Fisher’s exact test: P = 0.97; Fig. 2). Therefore, tree genus was not a predictor of aye-aye foraging preference in this study.
Fig. 2

Aye-aye foraging events on deadwood specimens (by genera) in Kianjavato, Madagascar (June–August 2013). The vertical bars represent the number of deadwood resources identified to each observed tree genus; genera are indicated in the figure key. The numbers of nonforaged and foraged trees are represented by dark gray and light gray bars, respectively. Of the total foraged N = 7 and nonforaged N = 156 trees, 1 and 26 trees, respectively, could not be identified to genus. As a result, the sample size for this analysis is N = 6 foraged and N = 130 nonforaged trees.

We did not observe any statistically significant preferences for either tree height (t-test: df = 7, P = 0.57) or diameter (t-test: df = 7, P = 0.36). When diameter measurements were subsampled to include only specimens in which diameter could be measured at breast height, results were still not statistically significant (t-test: df = 8.65, P = 0.11). However, although this result indicates that aye-ayes are not foraging on the majority of the deadwood resources of any particular size, visual inspection of the data suggests that there may be some avoidance of the tallest and largest diameter trees (Fig. 3). Specifically, although the mean height of foraged trees (6.83 m) is greater than that of nonforaged trees (5.48 m), there were no observations of foraging on trees >15.6 m. In contrast, 18 of 153 nonforaged trees (11.8%) were taller than 15.6 m. Similarly, even though the DBH difference was not statistically significant, the mean diameter for foraged specimens (18.81 cm) was 29% smaller than that of nonforaged specimens (26.62 cm). There were also no observations of foraging on trees with diameters >35.6 cm, whereas 21 of 150 nonforaged trees (14%) had greater diameters than the largest foraged deadwood resource.
Fig. 3

Boxplots of diameter and tree height for aye-aye-foraged vs. nonforaged deadwood resources in Kianjavato, Madagascar (June–August 2013). (A) Tree heights of foraged vs. nonforaged deadwood specimens. (B) Similar comparison for specimen diameter. For this study, we measured diameter at ca. 119 cm on all standing deadwood specimens and at the next closest available location for truncated or fragmented deadwood segments.

Deadwood Internal Structural Properties

The aforementioned results suggest that a different major aye-aye food resource, such as Canarium, may be the critical limiting resource in the aye-aye diet, or that the internal structural rather than external properties of deadwood might impact net nutritional gain from extractive foraging, or both. Unfortunately, as there were only two total foraged trees within our two 100 × 100 m plots and only five Canarium specimens (ESM Fig. S1), we were unable to statistically evaluate the Canarium proximity hypothesis in the present study. However, we were able to study variation in the internal properties of the deadwood resources by using the Fakopp Arborsonic 3D Acoustic Tomograph to estimate the velocity of sound traveling through the trees, a proxy for true density.

The exterior bark surfaces of 2 of the 7 (29%) foraged and 109 of the 156 (70%) nonforaged deadwood tree resources in our sample were too decayed to hold the Fakopp Acoustic Tomograph sensors tightly in place, which is necessary to obtain accurate sound velocity data. For each of the remaining deadwood trees in the sample, we conducted up to five cross-sectional Acoustic Tomograph scans spaced at heights ca. 5 cm apart and compiled velocity estimates at 1-cm intervals from the outer surface to the origin of the tree on four linear paths for each scan. Our final database for internal structure analysis included 224 scans from 47 nonforaged deadwood resources and 25 scans from 5 foraged resources (Fig. 4a).
Fig. 4

Internal structural properties of aye-aye foraged vs. nonforaged deadwood resources in Kianjavato, Madagascar (June–August 2013). (A) 2D renderings (acoustic tomographs) of nonforaged and foraged deadwood tree resources, as generated by the Fakopp Arborsonic Program. The renderings depict predicted specimen density based on the velocity (centimeters per second) readings of the scan. A representative rendering from one scan per nonforaged deadwood resource is shown. (B) Sound velocity (in cm/s) based on four values per scan for each 1-cm interval from depths of 1–10 cm, with the number of intervals for each scan dependent on the radius of the tree at the scanning location. The velocity estimates at discrete tree depths are compared here between foraged (in red) vs. nonforaged (in blue) deadwood tree specimens. (C) The Fakopp Arborsonic hardware. Thin, elongated probes are hammered into a deadwood specimen around its circumference at equivalent spacing intervals (the version of the machine shown has 8 probes; the version used in our study had 10 probes). At the tip of each probe is a sensor that detects sound waves traveling across the interior of the specimen. The external, flat surface of a probe is tapped with a hammer a minimum of three times. Sound propagates across the tree from tapping one probe. The other nine sensors measure the velocity of the sound waves as they reach each sensor tip. This process is repeated for each of the different probes. These velocity data are communicated to a laptop computer running the Fakopp Arborsonic Program via BlueTooth wireless software or a wired connection.

We observed that for the first 1–3 cm from the outer tree surfaces, the estimated velocity values were slightly but statistically significantly lower for foraged deadwood resources compared to nonforaged specimens (Fig. 4b). However, from 6 cm inward toward the center of the tree this pattern is reversed, with statistically significantly higher velocities for the foraged deadwood specimens. For example, at 6 cm, 7 cm, and 8 cm from the outer surface of the tree the average velocity estimates for foraged trees were 17%, 28%, and 40% greater, respectively, than those for nonforaged specimens (Welch two-sample t-tests: df = 73, P < 0.01; df = 71, P < 0.0001, and df = 38, P < 0.001; Fig. 4b).

Because we scanned foraged trees at the locations where traces occurred rather than strictly at DBH, the scanned heights were different for some of the foraged vs. nonforaged trees. However, when we restricted our comparison of foraged vs. nonforaged trees to only those scans that were taken from 80–120 cm heights (N = 3 foraged trees; N = 17 nonforaged trees), we observed similar results (ESM Fig. S2; t-tests for 6 cm, 7 cm, and 8 cm from tree outer surface: df = 68, P < 0.0001; df = 54, P < 0.0001; df = 20, P < 0.0001 respectively). Thus, while nonforaged trees have similar or slightly higher velocities on the 1- to 3-cm intervals compared to foraged trees, the deadwood resources selected by aye-ayes for insect larvae foraging tend to have relatively more intact interior cores compared to the overall sample of available deadwood.

The aforementioned t-tests of similarity for foraged vs. nonforaged velocity estimates likely violate the test’s assumption of independence among values. The density values from a given region of one deadwood specimen are likely related to the density values of different region within that same tree. Using a reduced dataset with a single mean velocity value for each cm interval per tree (see Methods) to repeat the analysis, only comparisons from the subset of 80–120 cm height scans remained statistically significant (all scans at 6 cm t-test: df = 4, P = 0.34; all scans at 7 cm t-test: df = 4, P = 0.15; subset of 80–120 cm scans at 6 cm t-test: df = 13, P = 0.05; subset of 80–120 cm scans at 7 cm t-test: df = 12, P < 0.01). Thus, our results should be considered tentative and preliminary until they can be replicated with a larger sample of foraged trees and, ideally, across additional aye-aye sites.

Discussion

We here investigated the distribution and properties of deadwood, a potentially important resource for the aye-aye diet that —if limited in some manner— could explain why aye-ayes maintain such large individual home ranges. However, we found that only a small fraction of all deadwood resources within aye-aye home ranges are accessed for insect larvae. Thus, either deadwood itself is not a limiting resource for aye-ayes, or particular properties of the deadwood are important, such that only subsets of all deadwood resources are limiting dietary factors for aye-ayes. If deadwood alone was a limiting resource, then a higher rate of deadwood foraging would likely have been observed, as aye-ayes would be expected to attempt to maximize nutrient gain relative to travel distance by foraging at all available dead trees. Alternatively, aye-ayes might attempt to maximize nutrient gain by repeatedly foraging on a singular resource with high larval content, once such a resource is discovered. Although we did observe indirect evidence of this behavior, with one heavily foraged specimen, this was a novel occurrence and does not change our overall conclusions.

We next compared external variables of foraged vs. nonforaged deadwood resources, beginning with taxonomy. We did not observe a preference toward any given deadwood genus. In her previous work at Nosy Mangabe, Sterling observed aye-ayes foraging on at least 6 families of insect larvae inhabiting a minimum of 29 different tree genera (Sterling 1993, 1994). Our observation that aye-ayes do not exhibit a strong preference for specific deadwood taxa is consistent with Sterling’s original result. Besides taxonomy, we also considered tree height and diameter as potential limiting factors. Within our sample, height and diameter were not statistically significantly different between foraged and nonforaged deadwood trees. However, it appears that there could be some preference for foraging on trees with smaller diameters, which in turn could be related to the ability to grip the tree properly during percussive and extractive foraging. Still, aye-ayes foraged on only a fraction of all available shorter and smaller diameter deadwood resources, suggesting that deadwood tree height and diameter alone are not major limiting factors for the aye-aye diet.

Finally, we compared the internal structures of foraged vs. nonforaged deadwood resources. We found that beyond 5 cm inward from the outer tree surface, foraged trees propagate sound more efficiently, i.e., have higher densities. Although this result is based on a limited number of observations and should thus be considered preliminary, it at least sparks an intriguing hypothesis related to aye-aye deadwood foraging ecology. Specifically, a previous study of aye-aye percussive foraging behavior reported that most larval excavation events occur within the first 1–3 cm from the outer surface of trees (Erickson 1994). Following our results, we tentatively hypothesize that the area behind the larval mines toward the tree center may have an important function in the aye-aye percussive foraging process. Specifically, this region could serve as a reflective “sounding board,” with denser wood in this region facilitating more precise acoustic reconstruction of the tree’s outer 1–3 cm, leading to more efficient foraging for insect larvae. Another factor that may contribute to deadwood resource selection could be independent of the tree structural properties required for aye-aye percussive foraging perceptual capability; namely, whether larval location varies by resource density. It may be that larvae preferentially occupy certain trees or particular regions therein based on deadwood interior or exterior structural properties. This will be an interesting topic for future research.

As only 10% or less of Madagascar remains forested, species with large home ranges are increasingly vulnerable to extinction (Mittermeier et al. 2010). Given their large and minimally overlapping (for females) home ranges, it is likely that aye-aye population densities are naturally very low (Mittermeier et al. 2010; Sterling 1993). Such factors, in conjunction with the aye-aye’s slow life history and relatively low genetic diversity, may make aye-ayes especially vulnerable to extinction (Catlett et al. 2010; Perry et al. 2012a, b, 2013; Schwitzer et al. 2013). An understanding of the role that limited food resources play in determining home range size may be crucial to our efforts of preserving this unique and Endangered primate (Schwitzer et al. 2013). Home range sizes and minimum conservation areas needed to maintain healthy populations of aye-ayes may vary as functions of the densities of deadwood, bamboo, Canarium, or other food items that comprise a smaller proportion of the aye-aye diet but may be seasonally or nutritionally critical, if those critical food resources are variable among different forest types within the species range, e.g., wet tropical rainforest vs. dry deciduous forest. Therefore, in addition to larger sample sizes of foraged trees, expanded analyses of Canarium resource locations, and the inclusion of other dietary resources and potential larval foraging from bamboo, future studies should include deadwood specimens measured at different times of the year, as rainfall, humidity, and temperature may all alter the decay process. Future studies could also investigate the relationship between deadwood density and the abundance of wood-boring insect larvae. As insect and tree species may also vary from region to region, this study should also be conducted in several different known aye-aye habitats to provide a more comprehensive view of the foraging preferences of this Endangered species.

Notes

Acknowledgments

We thank the Government of Madagascar for the permission to conduct research and the Madagascar Biodiversity Project (MBP) and its staff at the Kianjavato Ahmanson Field Station for facilitating this study, especially Razafindrahasy Alexander Théofrico (Frico), Kotozafy Gilbert André (Abanky), Randriambololona Stéphan Justin (Tofa), Fanoharanomenjanahary Hubert El-Phanger (Dadah), and Razafindrazefa Elysé Fortinand (Dagah). We also thank John Wickes, Peter Divos, and Akos Smuck of Fakopp Enterprise for their help and expertise regarding the ArborSonic 3D Acoustic Tomograph machinery and analysis program; James S. Solofondranohatra of the University of Antananarivo for his insight into aye-aye behavior and assistance in the field; and Zach Farris and Tim Sefzeck for contributing their compiled databases on the vernacular to scientific name translations for Malagasy tree species. We thank Steig Johnson and Nate Dominy for comments and discussion that helped shape this study; Logan Kistler, Martin Welker, Jeoren Smaers, Andrew Zamora, Rosemary Miller, and Annie Lin for insights or assistance with data analysis; Tim Ryan, Logan Kistler, Becki Coleman, and Stephen Johnson for comments on an earlier draft of this manuscript; and the constructive comments from two anonymous reviewers, the associate editor, and the editor of the journal that helped us to improve the paper. Funding was provided by the American Society of Primatologists, the Pennsylvania State University College of the Liberal Arts, The Pennsylvania State University Huck Institutes of the Life Sciences, and the benefactors of Pennsylvania State University Schreyer Honors College.

Supplementary material

10764_2016_9901_MOESM1_ESM.pdf (453 kb)
ESM 1(PDF 452 kb)

References

  1. Cartmill, M. (1974). Daubentonia, Dactylopsila, woodpeckers and klinorhynchy. In R. D. Martin, G. A. Doyle, & A. C. Walker (Eds.), Promisian biology (pp. 655–670). London: Gerald Duckworth and Co.Google Scholar
  2. Catlett, K. K., Schwartz, G. T., Godfrey, L. R., & Jungers, W. L. (2010). ‘Life history space’: a multivariate analysis of life history variation in extant and extinct Malagasy lemurs. American Journal of Physical Anthropology, 142(3), 391–404.CrossRefPubMedGoogle Scholar
  3. Coleman, M. N., & Ross, C. F. (2004). Primate auditory diversity and its influence on hearing performance. The Anatomical Record Part A: Discoveries in Molecular, Cellular, and Evolutionary Biology, 281(1), 1123–1137.CrossRefGoogle Scholar
  4. Di Bitetti, M. S. (2001). Home-range use by the tufted capuchin monkey (Cebus apella nigritus) in a subtropical rainforest of Argentina. Journal of Zoology, 253(1), 33–45.CrossRefGoogle Scholar
  5. Duckworth, J. W. (1993). Feeding damage left in bamboos, probably by aye-ayes (Daubentonia madagascariensis). International Journal of Primatology, 14(6), 927–931.CrossRefGoogle Scholar
  6. Erickson, C. J. (1994). Tap-scanning and extractive foraging in aye-ayes, Daubentonia madagascariensis. Folia Primatologica, 62(1–3), 125–135.CrossRefGoogle Scholar
  7. Farris, Z. J., Morelli, T. L., Sefczek, T., & Wright, P. C. (2011). Comparing aye-aye (Daubentonia madagascariensis) presence and distribution between degraded and non-degraded forest within Ranomafana National Park, Madagascar. Folia Primatologica, 82(2), 94–106.CrossRefGoogle Scholar
  8. Galdikas, B. M. (1988). Orangutan diet, range, and activity at Tanjung Puting, central Borneo. International Journal of Primatology, 9(1), 1–35.CrossRefGoogle Scholar
  9. Gerber, B. D., Arrigo-Nelson, S., Karpanty, S. M., Kotschwar, M., & Wright, P. C. (2012). Spatial ecology of the Endangered Milne-Edwards’ sifaka (Propithecus edwardsi): do logging and season affect home range and daily ranging patterns? International Journal of Primatology, 33(2), 305–321.CrossRefGoogle Scholar
  10. Hanya, G., & Chapman, C. (2012). Linking feeding ecology and population abundance: a review of food resource limitation on primates. Ecological Research, 28(2), 183–190.CrossRefGoogle Scholar
  11. Iwano, T., & Iwakawa, C. (1988). Feeding behavior of the aye-aye (Daubentonia madagascariensis) on nuts of ramy (Canarium madagascariensis). Folia Primatologica, 50(1–2), 139–142.Google Scholar
  12. Kaufman, J. A., Ahrens, E. T., Laidlaw, D. H., Zhang, S., & Allman, J. M. (2005). Anatomical analysis of an aye-aye brain (Daubentonia madagascariensis, Primates: Prosimii) combining histology, structural magnetic resonance imaging, and diffusion-tensor imaging. The Anatomical Record Part A: Discoveries in Molecular, Cellular, and Evolutionary Biology, 287(1), 1026–1037.CrossRefGoogle Scholar
  13. Mackinnon, J. (1974). The behaviour and ecology of wild orang-utans (Pongo pygmaeus). Animal Behaviour, 22(1), 3–74.CrossRefGoogle Scholar
  14. Manjaribe, C., Frasier, C. L., Bakolimalala, R., & Louis, E. E. (2013). Ecological restoration and reforestation of fragmented forests in Kianjavato, Madagascar. International Journal of Ecology, 2013, 1–12.CrossRefGoogle Scholar
  15. Mittermeier, R. A., Louis, E. E., Richardson, M., Schwitzer, C., Langrand, O., Rylands, A. B., Hawkins, F., Rajaobelina, S., Kappeler, P. M., Rasoloarison, R., Roos, C., & Mackinnon, J. (2010). Lemurs of Madagascar (3rd ed.). Bogota: Conservation International.Google Scholar
  16. Perry, G. H., Melsted, P., Marioni, J. C., Wang, Y., Bainer, R., Pickrell, J. K., Michelini, K., Zehr, S., Yoder, A. D., Stephens, M., Pritchard, J. K., & Gilad, Y. (2012a). Comparative RNA sequencing reveals substantial genetic variation in endangered primates. Genome Research, 22(4), 602–610.CrossRefPubMedPubMedCentralGoogle Scholar
  17. Perry, G. H., Reeves, D., Melsted, P., Ratan, A., Miller, W., Michelini, K., Louis, E. E., Pritchard, J. K., Mason, C. E., & Gilad, Y. (2012b). A genome sequence resource for the Aye-Aye (Daubentonia madagascariensis), a nocturnal lemur from Madagascar. Genome Biology and Evolution, 4(2), 126–135.CrossRefPubMedPubMedCentralGoogle Scholar
  18. Perry, G. H., Louis, E. E., Ratan, A., Bedoya-Reina, O. C., Burhans, R. C., Runhua, L., Johnson, S. E., Schuster, S. C., & Miller, W. (2013). Aye-aye population genomic analyses highlight an important center of endemism in northern Madagascar. Proceedings of the National Academy of Sciences of the United States of America, 110(15), 5823–5828.CrossRefPubMedPubMedCentralGoogle Scholar
  19. Pollock, J. I., Constable, D., Mittermeier, R. A., Ratsirarson, J., & Simons, H. (1985). A note on the diet and feeding behavior of the aye-aye Daubentonia madagascariensis. International Journal of Primatology, 6(4), 435–447.CrossRefGoogle Scholar
  20. Ramsier, M. A., & Dominy, N. J. (2012). Receiver bias and the acoustic ecology of aye-ayes (Daubentonia madagascariensis). Communicative and Integrative Biology, 5(6), 637–640.CrossRefPubMedPubMedCentralGoogle Scholar
  21. Schwitzer, C., Mittermeier, R. A., Davies, N., Johnson, S., Ratsimbazafy, J., Razafindramanana, J., Louis, E. E., & Rajaobelina, S. (Eds.). (2013). Lemurs of Madagascar: A strategy for their conservation 2013–2016. Bristol, UK: IUCN SSC Primate Specialist Group, Bristol Conservation and Science Foundation, and Conservation International.Google Scholar
  22. Sefczek, T., Farris, Z. J., & Wright, P. C. (2012). Aye-aye (Daubentonia madagascariensis) feeding strategies at Ranomafana National Park, Madagascar: an indirect sampling method. Folia Primatologica, 83(1), 1–10.CrossRefGoogle Scholar
  23. Simons, E. L. (1995). History, anatomy, subfossil record and management of Daubentonia madagascariensis. In L. Alterman (Ed.), Creatures of the dark: The nocturnal prosimians (pp. 133–140). New York: Plenum Press.CrossRefGoogle Scholar
  24. Singleton, I., & Van Schaik, C. P. (2001). Orangutan home range size and its determinants in a sumatran swamp forest. International Journal of Primatology, 22(6), 877–911.CrossRefGoogle Scholar
  25. Smith, R. J., & Jungers, W. L. (1997). Body mass in comparative primatology. Journal of Human Evolution, 32(6), 523–559.CrossRefPubMedGoogle Scholar
  26. Solofondranohatra, J. S. (2014). Ecoéthologie d’une femelle de Daubentonia madagascariensis dans la forêt de Kianjavato, sud-est de Madagascar. Mémoire de D.E.A. en sciences de la terre et de l’evolution. Spécialité: Primatologie, Université d’Antananarivo.Google Scholar
  27. Sterling, E. J. (1993). The behavioral ecology of the aye-aye on Nosy Mangabe, Madagascar. Ph.D. dissertation, Yale University.Google Scholar
  28. Sterling, E. J. (1994). Aye-ayes: specialists on structurally defended resources. Folia Primatologica, 62(1–3), 142–154.CrossRefGoogle Scholar
  29. Sterling, E. J. (2003). Daubentonia madagascariensis, aye-aye. In S. M. Goodman & J. P. Benstead (Eds.), The natural history of Madagascar (pp. 1348–1351). Chicago: University of Chicago Press.Google Scholar
  30. Sussman, R. W. (1991). Demography and social organization of free-ranging Lemur catta in the Beza Mahafaly Reserve, Madagascar. American Journal of Physical Anthropology, 84(1), 43–58.CrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media New York 2016

Authors and Affiliations

  • Katharine E. T. Thompson
    • 1
  • Richard J. Bankoff
    • 1
  • Edward E. LouisJr.
    • 2
  • George H. Perry
    • 1
    • 3
  1. 1.Department of AnthropologyPennsylvania State UniversityUniversity ParkUSA
  2. 2.Center for Conservation and ResearchOmaha’s Henry Doorly Zoo and AquariumOmahaUSA
  3. 3.Department of BiologyPennsylvania State UniversityUniversity ParkUSA

Personalised recommendations