Results

The means of the significantly different adaptations (arboreal, terrestrial, and aquatic substrate use; fruit-eating, grazing, fresh grass grazing and mixed feeding) show trends across habitats such that rainforests have high percentages of fruit-eating and arboreality with no grazing adaptations, whereas grasslands have almost no fruit-eating, no arboreal-ity and high percentages of grazing within their communities (Reed, 2008). These data can be seen in Fig. 14.1 where the means for each habitat group are presented. Similarly, the correspondence analysis of modern African sites and these significantly different adaptations produce clusters of similar habitats (Fig. 14.2). Forests and grasslands, at opposite ends of the x-axis are most distinct from woodland and bushland habitats. The latter habitats have less tree and bush cover than forests, but are not as open as grasslands. Thus the habitats align on the x-axis from the left with those that have more annual rainfall and are more closed (i.e., forests) to those habitats that are drier with more seasonal rainfall and open. This axis encompasses 74.15% of the difference from the expected chi-squared values and as such accounts for much of the variation among habitats. The y-axis, accounting for 14.45% of the variation among habitats, separates those sites in which there is abundant water (that is not necessarily based on local rains) from those that do not. The wet sites include the Linyanti Swamp (woodland), Kafue Flats (grassland), and the Okavango Delta (shrubland), all of which have a broad based habitats that are different from one another. They group towards the bottom of the graph with other sites that have lacustrine and/or fluvial systems, and as such, influence the types of mammals that utilize them.

Figure 14.3 shows the fossil localities in chronologically order from oldest to youngest with the grazing adaptation from each locality plotted. Grazing increases to 30% or

Fig. 14.1 The distribution of five significantly different adaptations across modern habitat types is depicted through means of each habitat type except desert where there is only one locality. Arboreal substrate use and fruit and leaf eating adaptations drop in abundance from forest through grasslands, whereas grazing adaptations rise across the same span. Aquatic substrate use and fresh grass grazing are not pictured as they separate all habitats on different criteria.

€——

X

❖ \ • \

\ \ S

A

\

-

- ir

/1 \

• —

~ —•- — _

Forests ) Closed Woodlands

Bushlands

6

Ecotones

■e o tn 9 a

-Frugi'/ory - -A- Grazing -- > ■ Terrestrial yn - Mixed Feeding

-Frugi'/ory - -A- Grazing -- > ■ Terrestrial yn - Mixed Feeding

Fig. 14.2 Correspondence analysis of seven significantly different There is a gradient from less seasonal environments on the left to adaptations among modern communities. Rainforests and grasslands greater seasonality and drier habitats on the right. Abbreviations as in are at opposite ends of the spectrum and are outlined in solid lines. Table 14.1.

Fig. 14.2 Correspondence analysis of seven significantly different There is a gradient from less seasonal environments on the left to adaptations among modern communities. Rainforests and grasslands greater seasonality and drier habitats on the right. Abbreviations as in are at opposite ends of the spectrum and are outlined in solid lines. Table 14.1.

Fig. 14.3 The distribution of grazing across the fossil sites used in this study. The fossil sites are arranged in chronological order and there is no consistent trend toward more grazing, although there is a jump in the percentage of grazers at ~2.0 Ma (Olduvai Bed I).

Fig. 14.3 The distribution of grazing across the fossil sites used in this study. The fossil sites are arranged in chronological order and there is no consistent trend toward more grazing, although there is a jump in the percentage of grazers at ~2.0 Ma (Olduvai Bed I).

greater in East African localities from approximately 2.0 Ma. This contrasts with all modern sites whose percentage of grazing mammals never exceeds 25%. Figure 14.4 shows the patterns fruit-eating, mixed feeding, and arboreality in the fossil sites. Overtime, there is no apparent trend in any one of these adaptations. If each region is separately examined in chronological order (Fig. 14.5) there are few direct trends within each. Although the end points in time for each location tend to have high grazing percentages and low percentages of fruit-eating and arboreality, none of the sites have a direct trend to that end. In relation to the hominins that have been recovered from the middle and late periods, there appears to be a larger change in mammal adaptations compared to what came before. That is, there is more change in mammal adaptations and habitats with the appearance of Homo erectus.

The correspondence analyses of each time period identified the habitat type of each fossil site. Figure 14.6 displays the early period centered around 3.0 Ma. The fossil sites at this time are positioned in the right half of the graph with seasonal rainfall and woodland bushland extant sites. This contrasts with pre 3.0 Ma when many fossil localities are in the midrange of the graph indicating more closed habitats (Reed and Rector, 2007). The mammals of Middle Lomekwi Member suggest a lacustrine or deltaic component as does the depositional environment. The Tulu Bor, Kada Hadar 1 and 2, and the Upper Lomekwi Members, are positioned with open woodlands although the depositional environment is also fluvial or lacustrine. The habitat of Makapansgat 3 is in the drier open woodland - scrub woodland range.

Figure 14.7 shows the placement of the middle time period sites. The WT-17000 site from which P. aethiopicus has been recovered, is reconstructed as bushland/medium density woodland. The fauna from Shungura C and F, as well as the Maakamitalu from Hadar indicates wooded grasslands, while Shungura D and E are drier shrubland habitats. The fauna from Shungura G and the Kalochoro Members cause these sites to be located to the left of the graph within modern grassland habitats. Sterkfontein 4 is positioned with the most arid of the modern habitats. The latest fossil sites from Koobi Fora and West Turkana (~1.6 Ma) fall in the range of modern habitats that include flood plains, swamps or lacustrine environs (Fig. 14.8). This is contrasted with the KBS and Upper Burgi Members of Koobi Fora, which are reconstructed as not as wet, and the Olduvai sites and Swartkrans 1, which are even more dry and open. The Olduvai mammals produce habitat reconstructions that are outside of the range of modern habitats, including the Serengeti Plains.

Finally, the cluster analysis provides a look at differences among species compositions at all sites (Fig. 14.9). Rather than grouping clusters by time, the first major break is between South and East Africa. This indicates that the fauna within each region is distinctly associated with the area. Sequential pan-African species turnovers, which would be represented by time clusters, are not evident. Within South African sites, each site is almost equidistant from the others, likely indicating similar amounts of species differences between these three time periods. Therefore, there appears to be significant species change every 500 ka in South Africa,

Fig. 14.4 The distribution of arboreality, fruit and leaf eating, and mixed feeding across the fossil sites. The fossil sites are arranged in chronological order and again there is no consistent trend towards less arboreality and fruit and leaf eating, nor an increase in mixed feeding.

Fig. 14.4 The distribution of arboreality, fruit and leaf eating, and mixed feeding across the fossil sites. The fossil sites are arranged in chronological order and again there is no consistent trend towards less arboreality and fruit and leaf eating, nor an increase in mixed feeding.

Fig. 14.5 The distribution of four significantly different adaptations across fossil sites arranged by formation or geographic region. In general there are slight trends in each area toward mammalian adaptation percentages that indicate more open, seasonal habitats especially at Koobi Fora and West Turkana.

100%

100%

West Turkana

Koobi Fora

Hadar

South Africa

Sbungura

Olduvai

A A A A

A A

A

A A

A

V

X

X X

X

X

X

X

• ;-T-; m , m I *

A

•- •

• î m . • . • T «

• î m î

I Arboreality ATerrestnality AFruitand Leaf Eating X Grazing

I Arboreality ATerrestnality AFruitand Leaf Eating X Grazing although without having sites in the missing time periods we cannot say if the changes were gradual, or if they involved migration or speciation events at specific intervals.

The East African sites are grouped by both region and time. First the early Hadar sites from which A. afarensis have been recovered are the most unique of this large cluster. This is probably because of its distance from the other more geographically restricted localities (Reed, 2008). The later A.L. 666 Homo cf. H. habilis site at Hadar (Makaamitalu), on the other hand, is positioned between the Olduvai and Turkana Basin Formations. The Olduvai Beds group further from this main cluster also likely due to distance from the greater Turkana Basin in northern Kenya and southern Ethiopia. It is not surprising that the Shungura

Fig. 14.6 Correspondence analysis of the adaptations with the modern localities and early time period fossil sites. Rainforests and grasslands are at opposite ends of the spectrum and outlined in solid lines. The fossil sites are distributed within medium density (wood, bush, or scrub), seasonal modern habitats.

Fig. 14.6 Correspondence analysis of the adaptations with the modern localities and early time period fossil sites. Rainforests and grasslands are at opposite ends of the spectrum and outlined in solid lines. The fossil sites are distributed within medium density (wood, bush, or scrub), seasonal modern habitats.

Fig. 14.7 Correspondence analysis of the adaptations with the modern localities and the middle time period fossil sites. Rainforests and grasslands are at opposite ends of the spectrum and outlined in solid lines. The fossil sites are generally distributed within low density, seasonal modern habitats such as scrub woodlands, shrublands, and grasslands. The Paranthropus aethiopicus site of WT17000 is more closed than all of the others of this time period. Shungura Members C and F, and the Makaamitalu from Hadar are depicted as habitats with greater proportions of wetlands or floodplains whereas the remaining sites are reconstructed as drier shrublands and grasslands.

sites group together irrespective of time within the large cluster because the region is proposed to have been a refugia against some of the major pan-African drying trends (Vrba, 1995). What is interesting is that the Shungura mammals more closely resemble those from the earlier Turkana Basin sites from the west side of the lake, while the sites later in time from the surrounding lake area (from ~1.8 Ma and younger) group together. In fact, all sites in this later group have had a fairly large species change from the previous time periods.

Was this article helpful?

0 0

Post a comment