In a series of five experiments with unimanual Go/No-Go tasks and intermixed Arabic numerical, i.e. numbers lower or higher than 5, and directional targets, i.e. arrows pointing to the left or to the right, we explored whether spatial codes used in isolation inherently evoke the left-to-right representation of number magnitudes, i.e. Space-to-Number congruency effect, and, vice-versa, whether number-magnitude codes used in isolation inherently evoke the conceptual activation of left/right spatial codes, i.e. Number-to-Space congruency effect. In Experiment 1 Go responses were provided based on instructions that activated only magnitude codes, e.g. "push only if the number is lower than 5 and whenever an arrow appears", or only spatial codes, e.g. "push only when an arrow points to the left and whenever a number appears". In Experiments 2-4, the same instructions were combined with the request of responding only to arrows in a specific colour. No fixed association was present between a specific arrow colour and a specific arrow direction. In Experiment 3, the direction of arrow-targets was kept fixed to favour the processing of arrow direction. In Experiment 4, an additional class of No-Go visual stimuli was included to heighten the focus of attention on numerical and directional arrow-targets and, in addition, only numbers 1, 2, 8 and 9 were used to force the contrast between small and large magnitude codes. The results of Experiments 1-4 highlighted no significant or reliable Space-to-Number congruency effect, e.g. faster RTs to numbers lower than 5 when participants attend to arrows pointing to the left, or Number-to-Space congruency effect, e.g. faster RTs to arrows pointing to the left when participants attend to numbers lower than 5. In Experiment 5 we confirmed that when spatial and number magnitude codes are used in conjunction, e.g. " push only when an arrow points left and when a number is lower than 5", a significant Space-to-Number congruency effect is found and we expanded this evidence to Number-to-Space congruency. Split-half testing showed that in Experiment 5 congruency effects were reliable and null-hypothesis significance testing showed that they were different from those found in all other experiments. We conclude that neither space codes used in isolation can elicit a spatial representation of number magnitudes nor number-magnitude codes used in isolation can trigger the activation of spatial codes. Thus, spatial and numerical codes must be used jointly to evoke reliable spatially organised mental number lines.