Seven logs from a large loquat (Eriobotrya japonica (Thunb.) Lindl.) tree were collected from a private residence in southern California (Granada Hills, CA, USA) The logs were similar in maximum diameter, length, and age. To verify candidate logs were active with drywood termites, using the methods from Lewis and Haverty [21], three 1-min recordings from the midpoint on the topside of the log were taken using a hand-held device (Tracker, Dunegan Engineering, Midland, TX, USA). All logs having at least 300 counts per minute were chosen to be included in the study. Five logs containing termites were selected to record AE activity and two logs containing no termites were chosen as controls to measure background AE activity from the surroundings.

The two control logs were placed into an oven (Isotemp model 655F, Fisher Scientific, Pittsburgh, PA, USA) at 105 °C for three days to kill any termites present in them. All seven logs had a subsurface sensor installed into their long center by drilling a 2.4-mm diameter hole and inserting the sensor probe 1.2 cm deep into wood. All log and sensor assignments were chosen randomly. A 3-m long cable from each of the seven sensors was connected into a port in the back of an AE smart device [29] and dedicated computer (Dell Corporation, Austin, TX, USA) that stored all of the data. The custom software used provided for twenty 3-minute recordings that were taken randomly among the seven sensors for each 60-minute period during the study. No sensor was used more than three times or less than two times during a 60-minute period, and the order of the readings was randomized. In addition, temperature and humidity (Omega Engineering, Stamford, CT, USA) were recorded with each 3-minute AE recording and saved to an electronic spreadsheet (Excel Corporation, Lubbock, TX, USA). A backup battery power supply (Back-ups, APC Corporate, W. Kingston, RI, USA) was also installed in the event of unexpected power outages. The entire AE data gathering and storage system was run twenty-four hours a day for almost 11 months (15 June 2008 to 15 May 2009). All logs, AE and temperature equipment were housed in a small wooden building at the University of California Richmond Field Station, Richmond, CA, USA. The building had five windows for natural light, and was without air conditioning, heaters, or insulation.

The hourly averages for AE ring down counts for each active and inactive log, as well as temperature and relative humidity, were plotted for the 11-month period. The mean daily activity pattern for I. minor in each of the five logs (sensors 1–5), represented by AE ring down counts, was graphed in relation to the mean temperature (°C) over a 24-hour period.

Seasonally, AE ring down counts displayed a non-linear pattern of increasing and decreasing values associated with temperature (Figures 1 and 2). Termite activity was highest during the warmer spring and summer months compared to winter. However, an increase in daytime temperature or a sudden heat wave, even in January and February 2009, resulted in a burst of termite activity (Figures 1 and 2). Humidity did not appear to have a significant impact on termite activity during the 11-month study.

Within a 24-hour day, AE ring down counts among the logs displayed a non-linear pattern of activity (Figure 3). These patterns were sinusoidal in shape for all sensors. Termite activity was lowest during the morning, increased in the afternoon, and peaked in late afternoon (1800), then declined until mid-morning. Based on the AE activity, the logs appeared to contain termite populations of different sizes: the smallest in log (sensor) 3, roughly three times larger in logs (sensors) 1, 4, and 5, and almost seven times larger in log (sensor) 2. The rise and fall in termite activity was correlated with temperature: activity increased as temperature rose (Figure 3). However, termite activity was apparently not significantly affected by relative humidity.

There are only a few reports available on seasonal or daily activity for I. minor as measured by AE technology. Using 100 workers contained in an artificially infested wooden block held at constant temperature and humidity, Lemaster et al. [22] presented AE event results over seven days from a single sensor. No statistically significant cycling or periodicity in AE measurements could be detected. The plot of termite activity appeared flat and ranged from just 100 to 200 AE events per hour.

There is considerable variance in the number of I. minor extracted from naturally infested logs and structural wood, and ranges from 7 to 2,943 [19,21,30-32]. Also clearly evident, larger colonies and infestations produce greater AE activity. And also clearly evident is that when drywood termites are allowed to search and forage for wood naturally under ambient conditions, their activity follows a cyclic pattern common to many terrestrial animals.

A second termite activity study using AE technology was conducted by Indrayani et al. [25] under laboratory conditions. Groups of 10 I. minor pseudergates were placed inside small wooden blocks to investigate the effects of varying temperature and humidity. The experiments lasted only 12 hours. The optimum temperature for peak AE activity was 30 °C. Reviewing the temperature and AE data from our study, termite activity increased with temperature, even during warm days during seasonally cold months (November 2008 to February 2009; Figures 1 and 2).

Knowledge on optimal times for drywood termite foraging could be important for termite inspections. Two species, I. minor and Cryptotermes brevis (Walker), are responsible for a majority of the damage caused by drywood termites in the United States [33-35]. The economic cost of control and repair of damage is second only to that of subterranean termites [34]. Traditional inspections are visual and are based on searches for damaged wood or pellets. The data from this study, as well as that in the Indrayani et al. [25] study, suggests searches for termite infestations could be enhanced by heating the wood to at least 25 °C prior to inspection to simulate foraging and feeding, even in winter. A more in-depth understanding the underlying mechanism that controls the cyclic pattern will require additional studies that include the exclusion of natural light, additional tests of naturally infested logs at constant temperature and humidity, and modifying the AE system collection hardware and software to sort out AE events or ring down counts due to the various behaviors including locomotion, feeding and foraging activity, and other vibrational behaviors.