We propose a statistical method for decomposition of contributions to iron production from various sources: supernovae Type II and the subpopulations of supernovae Type Ia -- prompt (their progenitors are short-lived stars of ages less then $sim$100
Myr) and tardy (whose progenitors are long-lived stars of ages $>$100 Myr). To do that, we develop a theory of oxygen and iron synthesis which takes into account the influence of spiral arms on amount of the above elements synthesized by both the supernovae Type II and prompt supernovae Ia. We solve this task without of any preliminary suppositions about the ratio among the portions of iron synthesized by the above sources. The relative portion of iron synthesized by tardy supernovae Ia for the life-time of the Galaxy is $sim$35 per cent (in the present ISM this portion is $sim$50 per cent). Correspondingly, the total portion of iron supplied to the disc by supernovae Type II and prompt supernovae Ia is $sim$65 per cent (in the present ISM this portion is $sim$50 per cent). The above result slightly depends on the adopted mass of oxygen and iron synthesized during one explosion of supernovae and the shape (bimodal or smooth) of the so-called Delay Time Distribution function. The portions of iron mass distributed between the short-lived supernovae are usually as follows: depending on the ejected masses of oxygen or iron during one supernovae Type II event the relative portion of iron, supplied to the Galactic disc for its age, varies in the range 12 - 32 per cent (in the present ISM 9-25 per cent); the portion supplied by prompt supernovae Ia to the Galactic disc is 33 - 53 per cent (in ISM 26 - 42 per cent).
Recently it has been proposed that there are two types of SN Ia progenitors -- short-lived and long-lived. On the basis of this idea, we develope a theory of a unified mechanism for the formation of the bimodal radial distribution of iron and oxygen
in the Galactic disc. The underlying cause for the formation of the fine structure of the radial abundance pattern is the influence of spiral arms, specifically, the combined effect of the corotation resonance and turbulent diffusion. From our modelling we conclude that to explain the bimodal radial distributions simultaneously for oxygen and iron and to obtain approximately equal total iron output from different types of supernovae, the mean ejected iron mass per supernova event should be the same as quoted in literature if maximum mass of stars, that eject heavy elements, is $50 M_{odot}$. For the upper mass limit of $70 M_{odot}$ the production of iron by a supernova II explosion should be increased by about 1.5 times.
