DYNAMICS OF ARTHROPOD FILIFORM HAIRS .2. MECHANICAL-PROPERTIES OF SPIDER TRICHOBOTHRIA (CUPIENNIUS-SALEI KEYS)

Adults of the wandering spider Cupiennius salei (Ctenidae) have 936 (/- 31 s.d.) trichobothria or filiform hairs on their legs and pedipalp s. This is the largest number of these air movement detectors recorded for a spider. The trichobothria are 100-1400 mum long and 5-15 mum wi de (diameter at base). Many of them are bent distally pointing towards the spider body. Their feathery surface increases drag forces and thu s mechanical sensitivity by enlarging the effective hair diameter. Typ ically, trichobothria are arranged in clusters of 2-30 hairs which inc rease in length towards the leg tip. The trichobothria's mechanical di rectionality is either isotropic or it exhibits a preference for air f low parallel or perpendicular (from lateral) to the long leg axis. The se differences are neither due to the distal bend of the hair nor to t he bilateral symmetry of the cuticular cup at the hair base but to the spring supporting the hair. Different directional properties may be c ombined in the same cluster of hairs. Trichobothria are tuned to best frequency ranges between 40 and 600 Hz depending on hair length. Becau se, with increasing hair length, absolute mechanical sensitivity chang es as well, the arrangement of hairs in a cluster provides for a fract ionation of both the intensity and frequency range of a stimulus, in a ddition, in some cases, to that of stimulus direction. Boundary layer thickness above the spider leg in oscillating airflow varies between a bout 2600 mum at 10 Hz and 600 mum at 950 Hz. It is well within the ra nge of hair lengths. In airflow perpendicular to the long leg axis par ticle velocity above the leg increases considerably as compared to the free field. The curved surface of the cuticular substrate has therefo re to be taken into account when calculating hair motion. The experime ntally measured properties of hair and air motion were also determined numerically using the theory developed in the companion paper (Humphr ey et al. Phil. Trans. R. Soc. Lond. B 340, 423-444 (1993)). There is good agreement between the two. Short hairs are as good or better velo city sensors as long hairs but more sensitive acceleration sensors. In agreement with most of our measurements optimal hair length is not la rger than boundary layer thickness at a hair's best frequency. Best fr equencies of hair deflection and of ratio a (maximum hair tip displace ment:air particle displacement) differ from each other. The highest me asured value for ratio a was 1.6. In only 22% of the cases hair tip di splacement exceeded air particle displacement. Hair motion is insensit ive to changes in hair mass as shown by the numerical comparison of a solid and a hollow hair.