I think genomic retrotransposable elements play a major role in evolution specialty under stress condition. Check the attached article

https://www.researchgate.net/publication/259912355_Environmental_Stress_Activation_of_Plant_LTR-Retrotransposons 

Article Environmental Stress Activation of Plant LTR-Retrotransposons

I am afraid you cannot separate the two of them. If environment changes, optimal combination of characters changes too and individuals (their alleles) closer to the optimal combination are going to be preferred.

Plus organisms are going to react with epigenetic changes, which can shift gene expression, or they make their mutation rate faster (e.g. in bacteria).

And last, but not least, changes in size can be direct reaction to nutrition changes in diet due to changes in the environment and any individual of the species would reach bigger/smaller size under one conditions then another just because there was more/less prey (i.e. nutrition and energy) when they grew up. (And changes in nutrition can possibly cause epigenetic changes...)

To sum up, it depends on what level of evolution are you focused in.

What do we understand as gradual evolution? Evolution occurs in all life forms from viruses to prokaryotes to diverse eukaryotes, such as invertebrates and vertebrates and plants. Prokaryotes evolve largely by means of horizontal gene transfer and natural selection, e.g. resistance to antibiotics. Vertebrates have many facilitators of evolution, such as whole genome duplications, point mutations, retroviral endogenisations, and  numerous others, together with natural selection. I take it that gradual evolution would mean adaptations rather than speciation or large innovations, such as the evolution of the placenta, for example . Intra-genomic potential* together with ecological pressure, both biotic and abiotic, and natural selection would be the main drivers of adaptation or gradual evolution in many eukaryotes. However, much evolution occurs in a punctuated equilibrium manner in eukaryotes. 

I hope this helps you,

Regards,

Keith

* Intra-genomic potential, as defined in the TE-Thrust hypothesis, is a continuum from realisable adaptive potential to realisable evolutionary potential.

There are balance between genotype and its optimal environment requirements to give its potential of gene expression for fitness, when the environment change the genotype face unsuitable conditions to gene expression and in this case natural selection will discard the unfavorable genotype by death or low fertility , while the new environment suitable for another genotype to be the normal , this refer to the importance of genetic diversity in evolution process , and the history of the earth told us many examples for the environment factor which control the process. 

A selection is single factor of gradual evolution (see, for example, famous book: Simpson, G. G. (1944) Tempo and Mode in Evolution. New York: Columbia Univ. Press).

I am not sure too that what you mean by "gene-from-gene". It is worth considering the phenotype is different than the genotype. You have yet to mention whether you are using a neutral mutation. If your gradual evolution uses this mechanism then another level of complexity arises with some coding and non-coding genes. Those have been valuable if PDGP and Cartesian Genetic Programming. Also this concept has been added to the tree-based GP, it is only a minority of researchers who uses it though. 

Non-coding genes can be transferred from one generation to another. Then they can become active to adapt to the environment. 

In the light of this debate please refer to:

http://www.anali-pazu.si/?q=sl/avtor/marko-vitas

and

http://rdcu.be/kojs

With kind regards,

Marko

Marko Vitas

Laze pri Borovnici 38,

Borovnica 1353,

Slovenia

Tel.: +386 41 362 111

[email protected]

http://www.anali-pazu.si/?q=sl/avtor/marko-vitas

http://rdcu.be/kojs

http://vitas-evolution.com/

http://www.dlib.si/details/URN:NBN:SI:DOC-UY4WVSCK/

https://plus.google.com/u/0/+MarkoVitasorigins/posts