however, certain peculiarities of position which are not characteristic of individual challenging. In the most common position the fish are about 1 inch apart, broadside to each other, head to tail, with their bodies flexed. The two together resemble parentheses. They usually remain in one place or drift slowly backward, forward, or sideways. Rare- This content downloaded from 205.208.116.024 on November 11, 2017 03:30:13 AM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). 156 JAMES C. BRADI)OCK AND ZORA I. BRADDOCK ly they revolve horizontally like a wheel while maintaining the same position with regard to each other. The position next in frequency resembles the first, except that the fish are head to head. They were never observed to revolve in this position. Other positions noted were a headon figure with the bodies flexed in opposite directions and a head-to-side arrangement, resembling the letter T with a tilted stem. Mutual challenges vary a great deal in duration. The maximum observed lasted 3 minutes and 30 seconds; the minimum less than 15 seconds. Also, they may be given in rapid succeschallenges during the period prior to fighting. If mutual challenges are treated as the individual challenges of two fish, i.e., added to the individual challenges of each participant, the maximum number of challenges per individual was 21, the minimum 1, and the mean 9. For contact pairs these figures were 35, 3, and 17, respectively. Some sort of challenging occurred in all experiments and was exhibited by every individual. The total number of all challenges was 605 (Table 3). These data contain no evidence of TABLE 4 FREQUENCIES OF CHALLENGING IN PERIOD PRIOR TO FIGHTING (35 EXPTS.) DURATION OF PERIOD (MINUTES) CHALLENGES 0-2 2-4 4-6 6-8 8-10 10-12 12-14* 16-18 18-20 i no. of individual challenges/ pair .................... 3 4 6 7 4 10 4 4 17 Sno. of mutual challenges/ pair.................... 1 5 5 6 12 10 7 8 6 i no. of all challenges/pair (indiv.+ 2X mutual+n)...... .5 13 15 17 28 29 18 20 29 * There were no periods having the time range of 14-16 minutes. sion, or the interval of time between them may be long. Behavior that starts as an individual challenge by one fish may become mutual challenging when the pair-mate counterchallenges and the instigator continues unchanged. As was the case with individual challenges, the mutual challenges continued throughout the period prior to fighting and were not more prevalent in one part of this period than another. The total number of mutual challenges for the 35 experiments was 199. They occurred in all the experiments. The maximum per contact pair was 15, the minimum 1, and the mean 6 (Table 3). It will be noted that the number of mutual challenges was approximately equal to the number of individual correlation between the duration of the period prior to fighting and the number of challenges. Indeed, certain of the shorter periods had more challenges than the long ones (Table 4). Counterchallenges are nothing more than individual challenges given in response to challenges by the pair-mate. They are included among the individual challenges mentioned above. Fifty-eight of the 207 individual challenges recorded for the periods prior to fighting were of this type. While challenging of one type or another was by far the most characteristic activity of the period prior to fighting, other activities occurred and should be recorded. Chasing was one of these. As This content downloaded from 205.208.116.024 on November 11, 2017 03:30:13 AM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). AGGRESSIVE BEHAVIOR OF FEMALE BETTAS 157 the name implies, this involves rapid pursuit around the aquarium of one pairmate by the other. The pursuer may or may not challenge while chasing. The total number of instances of this activity was 19. Ten of the seventy pair-mates chased, and this form of behavior occurred in 9 of the 35 fighting pairs. Considering only those individuals that actually chased, the maximum number of instances per individual was 3, the minimum 1, and the mean 2. Corresponding figures for contact pairs were 6, 1, and 2 (Table 3). Retreating occurred as a response to challenging, biting, chasing, or the approach of the pair-mate without signs of belligerence visible to the observer. It was not necessarily rapid, and it merely involved leaving the vicinity of the other individual. The total number of retreats during the period prior to fighting was 38. Fourteen of the seventy pair-mates retreated, and the activity occurred in 12 of the 35 experiments. Among those individuals that retreated, the maximum number per individual was 11, the minimum 1, and the mean 3. For contact pairs these figures were 15, 1, and 3 (Table 3). Thirty-three of the seventy pair-mates bit their opponents during the period prior to fighting. The total number of bites recorded for the period was 62. The activity occurred in 22 of the 35 experiments. Unlike all the other activities, biting was especially characteristico f the end of the period prior to fighting. In all but two of the experiments where it was observed, it occurred just before active fighting began. Based only upon the data recorded for those individuals that actually bit, the maximum per fish was 5, the minimum 1, and the mean 2. Corresponding figures for contact pairs were 6, 1, and 3 (Table 3). Exploring involves active swimming into various regions of the aquarium, poking the head into corners and the gravel on the bottom, and maintaining a generally alert posture. In contrast, the fish often swim about in an apparently aimless manner or rest quietly in one place. The total number of instances of exploratory activity was 28. Twenty-two of the pair-mates explored. This type of behavior was noted in 13 of the 35 experiments where fighting later occurred. Once again, considering only those cases where the type of behavior actually took place, the maximum number of instances of exploring per individual was 4, the minimum 1, and the mean 1. For contact pairs these figures were 7, 1, and 2, respectively. At intervals all individuals went to the surface for air. Bettas are labyrinth fishes, and this is part of their normal behavior. When this happened, all social activities ceased. No instances were recorded where a fish was challenged, bitten, or hindered in any other way while this was taking place. Reference to Table 3 indicates that challenging of all types was by far the most common activity during the period prior to fighting. On the basis of the number of participating pair-mates, biting was second, followed by exploring, retreating, and chasing in that order. However, the large variability among individuals and the small number of cases observed make it unwise to attempt to rank the relative frequencies of activities other than challenging. Various changes of color and pattern occurred during the period prior to fighting. Such changes are complicated and should be the subject of further investigation. One of the variables is the extent of saturation of the color of the body exclusive of stripes. Another is the saturation of the color of the median fins. The pectoral fins do not change This content downloaded from 205.208.116.024 on November 11, 2017 03:30:13 AM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). 158 JAMES C. BRADDOCK AND ZORA I. BRADDOCK color but remain colorless at all times. The pattern of the stripes which sometimes adorn the body also varies. These are always darker than the background color and may be longitudinal, dorsoventral, or plaid (a combination of the other two). Finally, the saturation of the color of the stripes themselves may vary. Each of these will now be discussed separately for the sake of clarity. The background color of the body may vary from pale to fully saturated. Changes may occur rapidly and involve passing through the intermediate stages. An individual engaged in contact-pair behavior may assume any phase of background body color, depending upon the nature of the behavior. The degree of saturation of the color of the median fins follows that of the background color of the body in a general way. Ordinarily, changes in one are accompanied by similar changes in the other. The saturation of fin color, however, is slightly greater than that of the body in all phases except that of full saturation. A few cases were noted in which there was little or no similarity between the color phases of body and fins. This suggests that they may be controlled by separate mechanisms. The patterns of striping are related to a certain extent to the phases of the background color of the body. Longitudinal stripes occur only when the latter is relatively pale. Dorsoventral stripes have been recorded in combination with all phases of saturation of background color and are especially characteristic of females in breeding condition. This is not their sole significance, however. The plaid effect has been noted only in combination with very pale background color, but it occurred so rarely that no conclusions should be drawn concerning the possibility that it may also accompany other phases. It may be a transitional stage between the two types of stripe pattern. Any of the striping patterns may change into any other without passing through a third, and the changes are sometimes very rapid. The color of the stripes themselves generally tends to become more intense as the background color lightens, but such is not always the case. The width of the stripes also varies. In the absence of stripes the background color of the body and the color of the fins may be of any degree of saturation. These facts would seem to indicate that striping patterns and color phases of the body and fins are all under the control of separate mechanisms.