Recently, a new lightning phenomena, termed needles, has been observed in both VHF and in optical along positive lightning leaders. They appear as small ($<$100 m) leader branches that undergo dielectric breakdown at regular intervals (called twinkles). Providing a coherent and consistent explanation for this phenomenon is challenging as each twinkle is a form of negative breakdown that propagates away from the positive leader. In this work we provide detailed observations of needles in VHF, observed during two lightning flashes. We show distributions of different needle properties, including twinkle propagation speeds, time between twinkles, and needle lengths, among others. We show a return stroke and multiple recoil leaders that quench needle activity. We also show that nearby needle activity does not seem to correlate together, and that needle twinkling can slow down by 10 to 30 percent per twinkle. We conclude by presenting possibilities for how the positive leader could induce negative propagation away from the positive channel, and we argue that twinkles can propagate like a stepped leader or like a recoil leader depending on the temperature of the needle, which implies that needle twinkles can probably propagate without emitting VHF.