Below many old continental flood basalt provinces typically there are contemporaneous sills, which are commonly quite thin. Some sills may show internal chilled margins that suggest repeated injection, but only once the previous sill had largely solidified and cooled. Much rarer are very thick intrusive bodies that have cooled sufficiently slowly to allow a layered complex to develop. It is generally accepted that they are the result of multiple intrusions that occurred prior to a significant degree of crystallization of the previous magma body. The Karoo Large Igneous Province and the Bushveld Complex are archetypal examples in South Africa of these processes. Both have volumes in excess of 106 km3. A consideration of the rate of cooling and crystallization of intrusions may provide an answer as to why the two different products result. Of 160 accurately measured sill thicknesses in the Karoo (from borecore information) relatively few exceed 50 m, and some may be composite. A 50 m thick basic sill will become 50% solidified in 10 years. The possibility of a second injection into the centre of that body in less than 10 years is generally unlikely. For a 100 m thick sill, the time period for the same degree of cooling is 37 years. Once the first sill has cooled by 100°C and become 50% crystalline, a second injection will chill, and form a distinct body within the first. It will not inflate the magma portion of the first sill. Average calculated emplacement rates for both the Karoo dolerites and the Bushveld Complex are far too slow for a layered complex to have developed. However, if there were three or perhaps four injections of magma about 100 m thick (or perhaps more than five injections of 50 m thickness) within 100 years into the same body, then the rate of cooling of that composite body could become sufficiently slow that it would act as a trap for any subsequent magma injection and a thick layered intrusion would result. Intensely pulsed volcanic episodicity in igneous provinces is beginning to be identified with increasing eruption intensity and decreasing time lapses, based on extremely precise measurements of palaeomagnetic directions. Hence, there need be no fundamentally different mechanism for the origin of the two different resultant igneous suites, merely a statistical peak in magma injection rate for a short period of time.