Indicadores de Gestión

Our field site and laboratory were at the Thomas Odhiambo Campus (Mbita Point, 0°25′S–0°30′S by 34°10′E–35°15′E, 1200m above sea level, mean annual temperature of 27°C) of the International Centre of Insect Physiology and Ecology (ICIPE) in western Kenya. All laboratory work was carried

out between 0800h and 1300h (laboratory pho- toperiod 12L:12D, lights on at 0700h). Standard spider-laboratory procedures were as in numerous previous studies (see Jackson & Hallas 1986; Cross et al. 2008) and only essential details are provided here.

We tested males (body length 5mm) and females (5mm) of E_. culicivora _{that had reached maturity}

(i.e., had undergone their final moult) 7–10 days earlier, had not mated and had not encountered any conspecific individuals since emerging from their egg sacs, as well as juveniles (i.e., immature indi- viduals that had moulted 8–12 days before used, were 3mm in body length, did not moult again in fewer than 10 days after used and, after moulting, were still immature). No spiders had been in contact with plants of any species before testing began, and no individual spider was tested more than once. The laboratory-rearing environment was “enriched” (spacious cages, mesh works of twigs within each cage), which has been shown to improve perform- ance of salticids in experiments (Carducci & Jakob 2000), and all spiders were maintained on a diet of chironomids and blood-fed Anopheles _{females three}

times a week (see Jackson et al. 2005).

A Y-shaped olfactometer (Fig. 2) was used to as- sess E. culicivora’s response to plant odours. With a Matheson FM-1000 flow meter, airflow was ad- justed to 1500ml/min. There was no evidence that this airflow setting impaired locomotion or had any adverse effects on E. culicivora’s behaviour. Air was pushed by a pump from a tap through two separate flow meters into two chambers, a stimulus chamber and a control chamber. Each chamber was a glass cube made from 5 mm thick glass (inner dimensions, 70 × 70 × 70mm), with a removable lid. There were two holes (diameter 20mm) in the cube that were opposite each other and each was plugged with a rubber stopper. There was a hole in each stopper through which a glass tube (diameter 4mm) passed. A nylon-netting screen over the stopper ensured that the test spider could not enter the chamber. New netting was used for each test. Air moved into and out of the stimulus chamber through the glass tubes to the stimulus arm, and independently from the control chamber to the control arm. Collectively, the two arms are referred to as the “choice arms”. Air moved from the two choice arms into the “test arm” (i.e., the stem of the Y).

The stimulus chamber contained a cutting from a plant and the control chamber was empty. Cuttings always included flowers, leaves and stems (i.e., we were interested in seeing whether E. culicivora

Cross & Jackson—Salticid-plant relationship 77

chooses plant odour, not just flower odour), and they came to half the height of the chamber (i.e., they did not rise above the level of the inflow and outflow holes in the stimulus chamber). Plant material was placed in the stimulus chamber 30min before each test. This 30 min period allowed air to circulate evenly and ensured that air pressure was comparable throughout the olfactometer. For each test, whether the stimulus chamber was on the left or right side of the olfactometer was decided at random.

A test spider was confined to a holding chamber (Fig. 2) at the far end of the test arm for 2min before testing began. A removable metal grill was fitted into a slit in the chamber roof, blocking access to the test arm from the holding chamber. The grill was lifted to start a test. Once the spider left the holding chamber, it was allowed 30min to make a choice (definition: entered a choice arm and remained there for 30 s). Spiders usually walked about actively in the olfactometer and we recorded which of the two arms it chose. As a precaution against the potential effects of traces left by spiders that had been tested previously, the olfactometer was dismantled and cleaned with 80% ethanol and then with distilled water between tests.

For this study, all data were analysed using chi- square tests for goodness of fit (null hypothesis: probability of making one of the two choices same as probability of making other choice) (see Howell

Fig. 1 Evarchaculicivora on plants. Juvenile and adult male on Ricinuscommunis. Adult female on Lantana camara.

Fig. 2 Olfactometer (not drawn to scale). Arrows indicate direction of airflow. Holding chamber (location of test spider

at start of test): length 25mm, internal diameter 25mm.

Start of test: test spider in holding chamber; grill removed, giving access to test arm, control arm and stimulus arm. Dimensions of test arm, control arm and stimulus arm:

length 90mm, internal diameter 20mm. Opaque barriers

New Zealand Journal of Zoology, 2009, Vol. 36 78

2002). For data analysis, individuals that failed to choose were ignored. For each test, N = 70. Voucher specimens of E. culicivora were depos- ited at the National Museums of Kenya (Nairobi), the Museum of Natural History (Wroclaw Univer- sity, Poland) and the Florida State Collection of Arthropods (Gainesville, Florida).

RESULTS

Evarchaculicivora males (χ² = 22.86, P < 0.001),

females (χ² = 8.23, P = 0.004) and juveniles (χ² = 4.63, P = 0.031) chose L. camara odour significantly more often than the control (Fig. 3A), and males (χ² = 6.91, P = 0.009), females (χ² = 14.63, P < 0.001) and juveniles (χ² = 5.71, P = 0.017) chose R. com-

munis odour significantly more often than the control

(Fig. 3B). For any given combination of spider sex- age class and plant species, the number of spiders that failed to choose was always less than 5%.