The term suddenly appears is more like created than evolved.
You’re getting hung up on improper semantics. In the absence to the contrary in the early stages of paleontology it was presumed there was likely to be comparatively long transition time periods between the differentiation of the populations. More recent evidence, however, has demonstrated how at least some populations can suddenly differentiate in only a few generations. The reason/s for the substantial differences in time periods for a population to differentiate is due to the interplay between the exact nature of the change in the population’s genome and how it affects the population’s ability to procreate and survive in its old and/or new ecological niche. Recent evidence has shown that populations under increased stress for survival induces biochemical activities which can in turn rapidly induce biochemical changes to the genome.
In the Andes Mountains of South America there is a population of Humans which have an incredibly high tolerance for the ingestion of the deadly poison, Arsenic. The groundwater in their area of habitation is heavily infused with Arsenic from the volcanic rock. Their ability to tolerate such a dangerous poison has been traced to a change in their genome which provides a biochemical means of more safely assimilating the Arsenic. In a paleolithic setting such a population of Humans could conceivably differentiate into another species of Humans given enough time and geographic isolation from other Human populations.
So, your assumption, “The changes should be gradual...,” is contrary to the factual evidence being discovered in many cases.