[adrian_b]

The early Hymenoptera were predatory, like wasps still are.

Bees, which feed on pollen and nectar from plant flowers, instead of catching other insects, have evolved much later from wasp-like Hymenoptera, after angiosperms were already widespread.

So the timeline fits into the story.

[xipho]

> were predatory ... ?

Were they? Sawflies are herbivores, and are at minimum share a common ancestor with those that were not herbivores.

https://www.sciencedirect.com/science/article/pii/S096098221...

[adrian_b]

From your link, the bees are nested within Crabronidae, most of which are predatory wasps, e.g. the beewolves.

So there is no doubt that the immediate ancestors of bees were predatory, like also their more distant ancestors that they share with ants and other kinds of wasps.

Hymenoptera has split early into two branches, one leading to wasps and the other to sawflies.

Most sawflies, as you say, feed on plants. So the question is what kind of food was used by the common ancestor of all Hymenoptera, given that from that ancestor there are two descendant groups, one predatory and one phytophagous (plant-eating).

There is no certain evidence for this, but there are arguments that it is much more likely that the common ancestor was predatory and the sawflies have evolved from that into plant-eating insects.

There are two arguments for this. One is that eating plants is more difficult for animals and all plant-eating animals had to develop special abilities in order to digest plants. Because of this there are a lot of examples among animal groups where the ancestors of a group were predatory, and only some of their descendants have succeeded to transition to feeding on plants.

There are much fewer examples of plant-eating animals that have become predatory again, because once some animals become able to eat plants, they normally have access to a much more abundant source of food than the predatory animals, so reverting to a predatory lifestyle would be disadvantageous. Humans are one of the few exceptions, because their ability to throw weapons at a distance has allowed them to become extremely successful hunters.

The second argument is that Hymenoptera form one of the two branches of Holometabola, the insects with complete metamorphosis. For the orders of insects that belong to the other branch of Holometabola, e.g. for beetles, wherever there is reasonable certainty about the lifestyle of the ancestors, they were also predatory, so it is more likely that the common ancestor of all Holometabola was also predatory.