Summary
Observations of territorial behaviour of male Nannothemis bella indicated that females only laid eggs on males' territories after copulating with the territory holder. Results of daily censuses of marked individuals on a study area suggested that a distinct subpopulation within the male population never retained a territory for more than 1 day on the study area. Such males tend to die or emigrate sooner than do males that hold a territory on the study area for more than 1 day at some time in their lives. Further observations of territorial behaviour showed that males holding territories for less than 1 day achieved copulations at significantly lower overall rates while on those territories than did males that retained territories for more than 1 day. We infer that all males attempt to hold territories for as long as possible, but under conditions of high population density a male must have a relatively high intrinsic "territory-holding ability" in order to hold a territory for more than 1 day. Census data indicated that males emerging late in the season were less likely to retain a territory for more than 1 day than males emerging earlier.
Methodology
The study area
The study area on Hebert Bog, near Chaffey's Locks in southern Ontario (44'30' N, 76'25' W) has been described by Hilder and Colgan (1985). A 90 X 50 m area of mat on this sphagnum bog was marked out with a grid of 4 x 4 m quadrats. There are numerous pools and areas of shallow water within this area, where female N. bella laid eggs and teneral individuals were observed emerging. There are also extensive relatively dry regions. The edge of the central area of open water (the "bog pond") forms one boundary of the study area; on the inland side, the area is bounded by a screen of trees and bushes (mostly black spruce and red maple), while the other two boundaries are arbitrarily marked. The study area is considered to be a typical and representative sample of N. bella breeding habitat. Adjacent regions of Hebert Bog are very similar to it. The behaviour of males observed on the study area is therefore assumed to be typical of their behaviour elsewhere.
Marking and census
Throughout the 1984 and 1985 seasons, a census was performed once a day, except when it rained all day. The entire study area was covered, an identical route being taken through the area each day. Teneral males (i.e. , < 1 day old) were caught, marked, and released, and the time and grid reference were noted. (Teneral males are easily recognized by their black colour and conspicuously shiny wings.) The times and grid references of all observations of previously marked males were also noted. In 1984, three searches were also made of other regions of the bog, up to a distance of ca. 30 m from the arbitrary boundaries of the study area. The marking technique was similar to that described by Connor (1971) and Hilder and Colgan (1985). Hilder and Colgan found evidence that the use of large bands of paint on the wings near the nodus and pterostigma might slightly reduce the individual's activity levels, so in this study only small spots of paint were placed on the leading edges of the wings, near the base and just inside the nodus. Using three different colours (orange, yellow, and white), up to 765 males could be marked. The mark could be read during a census without capturing or otherwise disturbing the individual. The census results were used to generate data on emergences, life-spans, and territoriality of males. The following simplifying assumptions were made: (i) When obtaining figures on the rates of emergence throughout the season, it had to be assumed that a roughly constant fraction of all males emerging in the study area each day were caught and marked. (ii) Males were presumed to have died or emigrated from the study area within 24 h of the last time they were recorded in a census. (iii) Males are described as "territorial" only if they were recorded more than once consecutively at the same location to within 2 m, in a region were the habitat was suitable for oviposition. Since censuses were only performed once a day, males holding a territory for less than 1 day were termed "nonterritorial." This arbitrary definition of territoriality is discussed further below. We assumed that any male holding a territory at the time of a census would be recorded. Wandering males, however, could be missed. (iv) The size of the study area is assumed to large relative to the distance that a male will typically wander in the course of its life. Phelan's observations described above suggest that this is the case, and further observations of our own tending to support this assumption are described below.
Observations of individual territoriality
In 1984, a total of 47 territorial males were observed for a number of 50-min periods between 0900 and 1800 Eastern Daylight Savings Time (solar noon at 1300), totalling 48 10 min of observation over 22 days in all, between the 8th and 36th days of the season. Numbers of fights, copulations, and ovipositions were recorded. In 1985, two 2 X 2 m quadrats within the study area were marked with lengths of string. Each quadrat contained pools of shallow water, with vegetation just below the surface, suitable for oviposition. Males holding territories within the quadrat were observed. These observations were carried out between the 13th and 39th days of the season. Observation was continuous between 1000 and 1600 on the days when there was >50% sunshine during this period. (This species shows little or no aggressive or sexual behaviour during cloudy weather.) On observation days, all unmarked adult males found within ca. 6 m of the quadrats before 1000 were marked with magenta paint spots on the wings to make them individually identifiable. Marking males later in the day appeared to disrupt their territorial behaviour, so unmarked males arriving the quadrats during the observation period had to be left unmarked until the following morning. The following data were recorded: days and times at which males acquired and lost territories within quadrats, and copulations and subsequent ovipositions obtained by males within the quadrats.