882 
regard the top of the haze layer (2-3 km) as the upper 
surface. This condition is confirmed by the sudden rise 
in humidity above this level, but when the tropical 
air has an oceanic source the surface of separation is 
ill-defined. Upper-air observations with. instruments 
more accurate than ordinary aircraft thermometers 
might make it possible to define the separation of 
upper and lower layers with more precision. 
When the surface air is equatorial, there is frequently 
a layer of trade or monsoon air before the overrunning 
easterlies are reached. The lower boundary of the trade 
or monsoon air is well marked by a rapid increase in 
wind speed with height, but the separation of trade 
or monsoon air and the easterlies is still ill-defined. 
Tt is quite clear that the day-to-day determination of 
the vertical structure of the air masses will not be pos- 
sible until observations are available in greater quan- 
tity and with more precision. 
FRONTAL PHENOMENA 
The question of the existence or nonexistence of true 
frontal discontinuities in the equatorial region is one 
of the most controversial issues of tropical meteorology. 
The majority of opinion, however, favours the existence 
of fronts, but the very fact that many writers disagree 
makes it important that factual evidence should be pro- 
duced. The acquisition of convincing evidence is ham- 
pered by the lack of close networks of upper-air sta- 
tions, and the main arguments put forward for the 
existence of fronts depend on the recognition of wind 
discontinuities and the acceptance of such discontinu- 
ities in wind as evidence of discontinuities in the air 
masses. Three distinct types of discontinuity are nor- 
mally recognised: the intertropical front, the meridional 
front, and subsidiary fronts. 
The intertropical front is regarded as being formed 
in the intertropical convergence zone, the zone com- 
pletely surrounding the earth in which the trade winds 
from the two hemispheres meet. In earlier days the 
zone was simply known as the doldrums and all the 
bad weather associated with the zone was regarded as 
due to the instability of the moist stagnant air which 
gave rise to convectional rain and thunderstorms. It 
is now more common to regard the zone as being one 
in which true frontal discontinuities between the north- 
ern and southern trades exist, interspersed with semi- 
stagnant equatorial air masses or doldrum areas. The 
boundary between the equatorial air masses and the 
trades will usually orient itself parallel to the direction 
of the active trade wind but with an increase or “‘surge”’ 
of the trade wind the relatively cold air will tend to 
undercut the equatorial air and give rise to frontal 
phenomena which are well marked and easy to iden- 
tify. The front between the trades, which is known as 
the intertropical front, forms most frequently at the 
times when the trade from the winter hemisphere is 
beginning its penetration into the summer hemisphere 
and when it is receding (z.e., just after the equinoxes). 
The advance of the trade across the equator is not a 
steady continuous process but is made up of surges 
TROPICAL METEOROLOGY 
and temporary retreats, each surge being accompanied 
by typical cold-frontal type phenomena; when the 
advancing trade dies down temporarily, the frontal ef- 
fect disappears to give place to a region of equatorial 
air, or the trade of the summer hemisphere temporarily 
reasserts itself. The retreat is a little more reeular, until 
the intertropical convergence zone has recrossed the 
equator, and is often accompanied by warm-front phe- 
nomena due to the flow of the freshening trade of the 
autumn hemisphere over the retreating trade, but more 
usually the activity of both trades at this period is small, 
and large areas of stagnant equatorial air appear on 
the synoptic chart. 
The meridional front is the front that forms between 
tropical air masses from two distinct anticyclonic cells 
in the same hemisphere. In the Southern Hemisphere, 
this takes place usually between a weak ESE/E stream 
from the more easterly of the two anticyclonic cells and 
a stronger SE/SSE stream from the more westerly cell. 
Because of the persistence of such cells during the win- 
ters of the respective hemispheres, these fronts are of a 
semipermanent character and are readily identifiable. 
In the summer of the hemisphere being considered, the 
portions of the cells are more variable with a conse- 
quent greater variation in the position of the meridional 
front. Where the air from the other hemisphere crosses 
the equator the meridional front crosses from the win- 
ter to the summer hemisphere, giving rise at its junc- 
tion with the intertropical convergence zone to a “‘triple 
point.”’ There is some evidence for the belief that this 
“triple point” is associated with the formation of trop- 
ical cyclones. 
Subsidiary fronts are of a more local character than 
the ones already described and are caused by local 
variations within an otherwise homogeneous air mass. 
The cause of the local variations may be orographic 
diversion of one portion of the air stream or the diver- 
sion of a portion of the air stream because of the break- 
down of geostrophic control as it enters the equatorial 
region. The diverted portion subsequently meets the 
undiverted portion again and gives rise to a local zone 
of convergence, which may be very limited in extent 
and may fluctuate rapidly in position. The movements 
of these zones are very difficult to follow since the cause 
of the movements is not necessarily related to the winds 
in the neighbourhood of the zone but is related to the 
previous history of the air mass in which the zone has 
formed. The slopes of the fronts have been determined 
in many cases, the criterion almost invariably being 
the height at which the wind change occurs at various 
points. Most of the values lie between 1 im 300 and 1 
in 500. There is no theoretical value for the slope with 
which comparison can be made because of the break- 
down of the geostrophic equations in the area and the 
absence of suitable equations to replace them. It has 
also not been determined satisfactorily to what height 
the discontinuities extend, the main difficulty here 
being the similarity between the upper layers of the 
trade winds and the lower layers of the overrunning 
easterlies. Where a front is formed from undercutting 
