If we were mean enough to flip a scorpion onto its' back (dorsal side) and took a look at the prosoma, we would see that all the legs join the body relatively close to one another (sometimes even fusing). In this region is a ventral plate known as a sternum. It is possibly made from 2 fused sternal plates (see discussion of the metosoma) but it may also be derived from the metasTomal plate seen in xiphosurans and eurypterids (I've written it in capitals as eurypterids have a sternum-like plate called a metasToma- it is neither mesosoma nor metasoma, confusing huh?!). The relative positions of the leg coxae and the sternum are among the more important features in the history of scorpion evolution.
Coxosternal region in a range of scorpions. A: Archaeoctonus B: Praearcturus C: The extant scorpion, Diplocentrus. The asterix marks the sternum. Images edited from Kjellesvig-Waering, 1986.
In the modern scorpions the sternum shape is always derived from a pentagonal form, though the terms 'triangular' and 'transverse' are applied to some of these variations. In early scorpions the shape of the sternum was much more variable, but based upon an oval shape (needle like, pentagonal, triangular and a number of other terms can be applied to these variations). Particularly elongate (needle-like) sterna are found in a handful of extant and extinct scorpions (Praearcturus and Troglocormus springing instantly to mind.), scientists aren't too sure of the significance of this but it may be to aid spermatophore placement.
Over time, the sternum moved posteriorally, as first the anterior pair of coxae moved in front and fused, and then the second (and occassionally 3rd). During this progression lobes began to develop on the 1st (and much later, 2nd) pair of coxae. Known as coxapophyses, Stoermeroscorpio and Proscorpius are the earliest scorpions to show development of these coxal lobes. The 1st coxapophysis was well developed by the time of Praearcturus (whose 1st two pairs of coxae meet infront of the sternum; see 'B' in the figure above) and by the Mesoscorpionina the second coxae have maxillary lobes (just another term for the coxapophyses), which may, or may not, extend to the 1st maxillary lobes (as in 'C' in the figure above).
The trends in the evolution of the coxosternal (also known as the sternocoxal-) region are related to the change from particulate to liquid feeding and improved terrestrial movement. In Proscorpius and Waeringoscorpio the coxae may have been gnathobasic (a term meaning they might have been used in the crushing/tearing of food) as none of the four coxae appear fused. Although not occurring as a linear transgression to the more advanced state, the development of an oral tube made by the coxapophyses is indicative of improved liquid feeding (where the oral tube acts like a straw in sucking up the externally digested prey). Whilst it cannot be said with any certainty, it is possible that any anterior lobe may have helped with this. Likewise, any fusing of the coxae may have strengthened the scorpions ability for terrestrial locomotion.