But the reasons behind the rural-urban divide are unclear, although smaller rolls and composite classes may be factors. If all the school’s pupils are in the playground at the same time, there may be less segregation and more informal play.

Headteachers have been told that all schools must become health-promoting by 2007 and the survey’s findings are likely to be included in national advice on how to encourage children to be more physically active and fit.

The research, commissioned by Grounds for Learning and its project partners, Play Scotland and Sportscotland, found that 45 per cent of schools reported that all pupils were engaged in active play during breaks, while 31 per cent said that almost all their pupils took part.

Behind these headline figures there were significant differences across sectors, with activity levels declining dramatically by the time pupils reach secondary.

The research, which will be published in a final report in the spring, found that Argyll and Bute topped the activity league table with 69 per cent of all responding schools reporting that all children took part in active play at break time - well above the average.

By contrast, Glasgow reported only 30 per cent, West Lothian 27 per cent and East Renfrewshire 21 per cent.

A rural-urban divide also emerged in nursery schools, with higher proportions of schools from rural areas reporting that all children were engaged in active play in school grounds.

Compared to a nursery school average of 45 per cent, the proportion of schools in which all children take part in active play in the school grounds is higher in Highland (74 per cent), Dumfries and Galloway (75 per cent), Argyll and Bute, Moray and Scottish Borders (all 83 per cent).

In comparison, the authorities which showed lower than average rates for nursery pupils engaged in active play were concentrated in the central belt: West Lothian (12 per cent), Inverclyde (25 per cent) and Glasgow (26 per cent).

Small schools, in both the primary and secondary sectors, appear to have youngsters who are more physically active. Forty-one per cent of the smallest secondary schools - classed as having 200 or fewer pupils - reported that almost all pupils engaged in active play compared to 9 per cent of the largest secondaries (more than 1,000 pupils).

While the proportion of pupils engaged in active play was not associated with the levels of sports pitch ownership, there was an association between active play and provision of play activities.

The researchers, however, point out: “It is important to critically appreciate the nature of break time activity and not to interpret active and non-active play as a binary of ‘good’ and ‘bad’ behaviour.

“Non-active social interaction and friendship building is an important school ground activity, particularly for older children. Indeed, some forms of active playground play - the dominance of boys playing football - may even be viewed as a problem of the playground and a problem for society.”

While 97 per cent of primary schools, 92 per cent of secondaries, 79 per cent of special schools and 52 per cent of nursery schools have hard surface playgrounds, nine out of 10 respondents said there should be more variety.

Penny Martin, senior development manager at Grounds for Learning, said:

“Lots of variation and textures and surfaces stimulate children to run and move more than a flat undeveloped Tarmac area does.”

Schools provided spaces for informal activity as well as formal sports activity.