PROVIDENCE, RHODE ISLAND. — Providence, Rhode Island -- On Aug. 19, a colleague from Czechoslovakia and I were discussing the breaking news of the coup attempt in the Soviet Union. The reports of tanks in the streets were a bitter reminder to him that it had been 23 years almost to the day since the armored forces of the Warsaw Pact had invaded Czechoslovakia.
August seems to be a busy month for international violence: Last August, Iraq invaded Kuwait; in August 1914, Germany invaded Belgium and France; and in August 1939, Hitler and his generals plotted the September 1 invasion of Poland.
President Bush has every reason to ask, ''What is it about August?'' when, for the second year in a row, his vacation was interrupted by major international developments.
The question of whether there is some inherent connection between the time of the year and the outbreak of war has been addressed in the latest issue of the journal Nature by four Israeli authors, Schreiber, Avissar, Tzahor and Grisaru.
They studied the starting dates of more than 2,000 wars over the last 3,500 years and found that the outbreaks of wars occurring between latitudes 30 -- 60 degrees north (approximately our own temperate zone) rise steadily from a low in January to a peak in August, after which there is once again a steady decline to the January low.
The graph they present is a striking one since there is approximately a four-fold increase from the number of wars begun in January and February to those begun in June, July and August.
The Israeli authors then compare these results with the opening dates of wars between latitudes 30 -- 60 degrees south in the Southern Hemisphere. Here the pattern is reversed: The peak is in December, January and February, followed by steady decline to June and July.
In both the Northern and Southern hemispheres, the curve illustrating the numbers of outbreaks of wars per month is paralleled to the curve documenting the increase in day length.
Naturally, these correlations raise the question of what happens near the equator, where the day length shows little fluctuation over the course of the year. The answer is that wars are equally likely to start in January and December as in July and August.
The annual rhythm in the onset of wars strikes a responsive chord with biologists who have documented many examples of annual rhythms in temperate regions. These natural rhythms, based on the changing light light cycles of the year, have been shown to frequently govern important activities in animals, such as hibernation and breeding.
The authors of the correspondence in Nature restrict themselves to pointing out the correlation between increasing day length, irrational aggressiveness and the onset of wars. They speculate that ''elongation of the daily photoperiod may induce increases in affective aggressiveness.''
They do not, however, consider whether hormones might be involved. (One study, conducted in the Northern Hemisphere, showed that there is an annual rhythm in the concentration of the hormone testosterone in men's blood, with the highest level being present from August to November; this hormone has been associated with aggressive behavior in some male animals.)
Unfortunately, the authors do not discuss an alternative explanation: The logistics of war may govern their starting dates.
An aggressive leader is unlikely to initiate a war in the cold of winter or in the mud of spring when movement and maintenance of armed forces is difficult. Similarly, an agrarian is presumably less likely to launch a war during the spring and autumn when male labor is needed for sowing and harvesting.
In spite of these qualifications, the Nature correspondence deserves careful consideration. Its timely observations should alert world leaders to the dangers inherent in the late summer season.
Perhaps in the future, President Bush will be less surprised at the interruptions to his August vacation.
Peter Heywood is a professor of biology at Brown University.