Necessary, because if this were not true — if using the phases of the moon does NOT usually lead gardeners to plant later than those planting at the first warm spell — then the premises give us no reason to think that gardeners who plant by phases of the moon are less likely to lose plants to a frost.

Not necessary, because the author’s assumption relates to how the phases of the moon influence when a gardener plants. It’s not about any relationship between the moon phases and actual weather/climate conditions on Earth. If the phases of the moon do NOT affect whether a frost follows the first warm spell, they can still influence when gardeners first plant in the year.

Not necessary, because even if the moon-gardeners use the same kinds of plants, the moon-gardeners could plant later in the year, which is what contributes to their lower chance of losing plants to frost.

Not necessary, because even if gardeners can improve their results without understanding how their methods work, that’s consistent with the reasoning. What matters is that moon-gardeners plant later in the year; whether they understand why this is more successful doesn’t change whether they plant later in the year.

The argument doesn’t concern professional gardeners, only amateur gardeners.