## Abstract

Recombination in the ionized layers is discussed. It is pointed out that the results of recent work make the ionic recombination theory very difficult to maintain. Consideration is therefore given to two alternatives, the dust recombination theory and the molecular recombination theory. It is concluded that of these only the latter is at all promising. In the E layer the recombination process would be O$_{2}^{+}$ + e $\rightarrow $ O$^{\prime}$ + O$^{\prime \prime}$, while in F$_{1}$ the effective reactions would be $ \matrix \text{O}^{+}+XY\rightarrow XY^{+}+\text{O} \\ XY^{+}+e\rightarrow X^{\prime}+Y^{\prime} \endmatrix \bigg\}$, $\ \ \text{N}_{2}^{+}+e\rightarrow \text{N}^{\prime}+\text{N}^{\prime \prime}$, and $ \matrix \text{N}_{2}^{+}+XY\rightarrow XY^{+}+\text{N}_{2} \\ XY^{+}+e\rightarrow X^{\prime}+Y^{\prime} \endmatrix \bigg\}$. The molecule XY possesses a lower ionization potential than O and need form only a very small fraction of the upper atmospheric content at the altitudes concerned. It is not identified but may be O$_{2}$ or NO. The reaction involving O$^{+}$ must be introduced, as otherwise the concentration of this ion would build up to impossibly large values. In F$_{2}$ the same reactions would be supposed to occur as in F$_{1}$, but the reduced amount of XY would lead to a lower and pressure-dependent recombination coefficient. Confirmation of the theory must await proper determination of the reaction rates.