1 Introduction

Solar activity indications do not occur evenly on the solar disk. It is well known that many types of solar phenomena exhibit some N-S asymmetry distributions (Reid 1968; Hansen & Hansen 1975; Roy 1977; Swinson et al. 1986; Verma 1987; Verma et al. 1987; Garcia 1990; Oliver & Ballester 1994; Atac & Ozguc 1996; Joshi 1995; Heras et al. 1996). One of the first studies concerning the N-S asymmetry was made by Newton & Milson (1955), who studied the distribution of the yearly values of sunspot areas from 1874 to 1955. Later on, different studies of N-S asymmetry have been made based on different types of solar phenomena, relative sunspot numbers, sunspot areas, sunspot magnetic classes, sudden disappearances of solar filaments, type II radio bursts, white-light flares, gamma-ray bursts, hard X-ray bursts, events of coronal mass ejection, and X-ray flares (see the above references). These investigations have demonstrated the existence of a N-S asymmetry, its quasi periodic behavior of 11 years and its correlations with some phase difference with other physical phenomena. Although the existence of N-S asymmetry in solar activity is generally well established, it is not so well interpreted (Carbonell et al. 1993). Garcia (1990) studied the N-S asymmetry of large flares based on X-ray observations from GOES satellites during Solar Cycles 20 and 21. He shows that the spatial distribution of flares varies with solar cycle such as that the preponderance of flares occurs in the North in the early part of the cycle and moves south as the cycle progresses. Bai (1987, 1988) remarked that active regions producing flares and active regions with low activity could originate from different levels of the convection zone. The first class of active regions could come from superactive zones with rotation periods substantially shorter than the Carrington rotation period and comparable to that of large scale magnetic polarity patterns (McIntosh 1981). McIntosh et al. (1985) proposed that this large scale pattern is the result of giant convective cells. The association of these observations leads Garcia to think that flares originate in active regions anchored deep in the convection zone, may be at the boundary of giant cells. So there is, may be, a relationship between the asymmetries and giant convective cells. A global convection pattern was recently discovered by analysing the data of MDI aboard SOHO (Kosovitchev et al. 1997) and large torsional zones have been confirmed. The existence of the torsional oscillations could be related to the rolls defined by Ribes et al. (1993) using young spots as tracers. They show that the rolls parallel to the equatorial plane describe well the meridian circulation. The number of such convective rolls changes according to the phase of the solar cycle from 3 at the onset of the cycle to 1 at the decay of the cycle. Because the sun is a non rigid body in rotation, the meridian circulation has as a consequence the differential rotation of the surface due to the transport of momentum and magnetic activity. All these works would give a new view on the solar dynamo phenomenon.

On the other hand, Verma (1987) who was working on 3 Solar Cycles 19, 20 and 21 with different indicators found that asymmetries in all phenomena prevailed in the north during cycles 19 and 20 and in the south in cycle 21. This would indicate that superimposed on the well observed periodicities of 11, 22 years a much longer cycle of unknown duration does exist.

The existence of a E-W asymmetry is more controversial. During short periods (1978-1980) heliolongitudinal distribution of intense flares shows an important asymmetry towards the east and it has been related to the large number of interplanetary shocks observed in the eastern hemisphere (Hewish & Bravo 1986). Horas et al. (1990) found a pronounced asymmetry of flares and subflares between 1976-1985, although the asymmetry was much more important during periods close to the minimum of solar activity while around the maximum it was very small. Joshi (1995) using H$\alpha$ flares found no E-W asymmetry for Solar Cycle 21 and a small asymmetry during the Solar Cycle 22. There is no obvious reason why the W-E asymetry should exist over a long period. Some ideas have been proposed to explain why we could find a E-W asymmetry. Some gravitational effects with the interaction of Jupiter have been suggested in the past but it does not seem very promising (Kleczek private discussion). An other possibility is the influence over a few solar rotations of the transit of active flaring regions on the solar disk (Heras et al. 1990). In fact preferred heliolongitudes for solar activity have recently been pointed out by Bai (1987, 1988). Finally an other idea would be due to the pure geometry of the magnetic field in the sunspots. According to Bai (1988) and McIntosh (see above) the active regions are anchored deep in the convective zone. In these regions there are strong magnetic shears. The magnetic field lines when they emerge at the solar surface would have some inclination. So the area of sunspots could have different sizes in the western and in the eastern hemispheres according to the inclination of the field lines. If the angle is similar to this of the prominences, inclination of 15 degrees with the vertical towards the west (Tandberg-Hanssen 1974), the area of the sunspots should be larger on the eastern than on the western side. Again this is a topics of great importance with the recent development of seismic approachs with SOHO/MDI (Braun et al. 1993).

In this paper, we have used the data of the X-ray flares ($M\geq 1$) during the maximum period of the 22nd solar cycle in order to investigate the behavior of asymmetries: W-E asymmetry as well as N-S asymmetry in order to complement the work done by Garcia (1990) for the previous solar cycles and to compare with the results of Joshi (1995) and Heras et al. (1990) concerning Solar Cycle 22 but using other indicators. We also study distributions of the flares with respect to the longitude. We confirm our results by using probability laws in order to check if they could be obtained quite by chance.