We report the results of a comprehensive study of the relationship between galaxy size , stellar mass and specific star-formation rate ( sSFR ) at redshifts 1.3 < z < 1.5 .
Based on a mass complete ( M _ { \star } \geq 6 \times 10 ^ { 10 } \hbox { $ \thinspace M _ { \odot } $ } ) , spectroscopic sample from the UKIDSS Ultra-deep Survey ( UDS ) , with accurate stellar-mass measurements derived from spectro-photometric fitting , we find that at z \simeq 1.4 the location of massive galaxies on the size-mass plane is determined primarily by their sSFR .
At this epoch we find that massive galaxies which are passive ( sSFR \leq 0.1 Gyr ^ { -1 } ) follow a tight size-mass relation , with half-light radii a factor f _ { g } = 2.4 \pm { 0.2 } smaller than their local counterparts .
Moreover , amongst the passive sub-sample we find no evidence that the off-set from the local size-mass relation is a function of stellar population age .
In contrast , we find that massive star-forming galaxies at this epoch lie closer to the local late-type size-mass relation and are only a factor f _ { g } = 1.6 \pm { 0.2 } smaller than observed locally .
Based on a sub-sample with dynamical mass estimates , which consists of both passive and star-forming objects , we also derive an independent estimate of f _ { g } = ~ { } 2.3 \pm { 0.3 } for the typical growth in half-light radius between z \simeq 1.4 and the present day .
Focusing on the passive sub-sample , we conclude that to produce the necessary evolution predominantly via major mergers would require an unfeasible number of merger events and over populate the high-mass end of the local stellar mass function .
In contrast , we find that a scenario in which mass accretion is dominated by minor mergers can comfortably produce the necessary evolution , whereby an increase in stellar mass of only a factor of \simeq 2 , accompanied by size growth of a factor of \simeq 3.5 , is required to reconcile the size-mass relation at z \simeq 1.4 with that observed locally .
Finally , we note that a significant fraction ( 44 \% \pm 12 \% ) of the passive galaxies in our sample have a disk-like morphology , providing additional evidence that separate physical processes are responsible for the quenching of star-formation and morphological transformation in massive galaxies .