Narrative Description

Although many carbonaceous compressions have functional morphologies generally consistent with algal interpretation, their exact phylogenetic affinities are poorly resolved because of pervasive morphological convergence among algae. Possible exceptions include Miaohephyton bifurcatum and Beltanelliformis brunsae from the Doushantuo Formation; these have been compared, respectively, with fucalean brown algae and the coenocytic green alga Derbesia (Xiao et al, 1998a; Xiao et al., 2002). In addition, several microscopic compressions recovered from Proterozoic shales using palynological method are phylogenetically resolved. For example, Proterocladus major from the ~750 Ma Svanbergfjellet Formation in Spitsbergen has been interpreted as a clodophoran green alga (Butterfield et al, 1994). Palaeovaucheria clavata from the ~1000 Ma Lakhanda Group in southeastern Siberia and Jacutianema solubila from the Svanbergfjellet Formation are both interpreted as xanthophyte algae (Hermann, 1990; Butterfield, 2004). Finally, the silicified microfossil Bangiomorpha pubescens from the ~1200 Ma Hunting Formation in Arctic Canada has been interpreted as a bangiophyte red alga (Butterfield, 2000), and several phosphatized algae from the Ediacaran Doushantuo Formation have been interpreted as florideophyte red algae (Xiao et al., 2004). These fossils indicate that major algal clades diverged no later than the early Neoproterozoic (Knoll, 1992; Porter, 2004).

However, clade divergence needs not be temporally coupled with morphological, ecological, and taxonomic diversification. Therefore, it is useful to independently characterize important morphological innovations in macroalgal history. We begin by tabulating the temporal distribution of some important macroalgal morphologies (Table 1), followed by a brief summary of macroalgal morphologies in the Proterozoic.

Paleoproterozoic and Mesoproterozoic macroalgae are mostly spherical, ellipsoidal, tomaculate, or cylindrical forms. Carbonaceous compressions similar to Chuaria, Ellipsophysa, and Tawuia are known from the 1800-1700 Ma Changzhougou and Chuanlinggou formations in North China (Hofmann and Chen, 1981; Lu and Li, 1991; Zhu et al., 2000; Wan et al., 2003), although those from the Changzhougou Formation have recently been characterized as pseudofossils (Lamb et al, 2005). Grypania and Grypania-like fossils have been reported from the ~1900 Ma Negaunee Iron-Formation of Michigan (Han and Runnegar, 1992; Schneider et al., 2002), the Mesoproterozoic Rohtas Formation of central India (Kumar, 1995; Rasmussen et al, 2002; Ray et al., 2002), and the ~1400 Ma Gaoyuzhuang Formation in North China and the Greyson Shale in Montana (Walter et al., 1990); the Indian Grypania specimens are distinct in bearing transverse annulations. Abundant carbonaceous compressions occur in the ~1700 Ma Tuanshanzi Formation in the Jixian area (Hofmann and Chen, 1981; Yan, 1995; Zhu and Chen, 1995; Yan and Liu, 1997). Some of the Tuanshanzi fossils have been interpreted as macroalgae with holdfast-stipe-blade differentiation, but their variable morphologies appear to suggest that some of them may be fragmented algal mats. However, Tawuia-like fossils from the Mesoproterozoic Suket Shale in central India do appear to have simple discoidal holdfasts (Kumar, 2001).

Table 1. Temporal distribution of important macroalgal features (+: presence; ?: possible presence).

Paleoproterozoic (25OO-16OO Ma)

Mesoproterozoic (16OO-1OOO Ma)

Early

Neoproterozoic (1OOO-75O Ma)

Ediacaran (635-542 Ma)

Thallus Morphologies

Spherical

+

+

+

+

Ellipsoidal

+

+

+

+

Tomaculate

+

+

+

+

Cylindrical

+

+

+

+

Conical

+

Fan-shaped

+

Thallus Differentiation

Holdfast

?

+

+

+

present

Discoidal

+

+

+

holdfast

Rhizoidal

+

holdfast

Stipe

?

+

+

Blade

?

+

Other Features

Transverse

+

+

+

annulation

Dichotomous

+

Branching

Monopodial branching

+

Apical

+

meristem

Early Neoproterozoic macroalgal assemblages continued to be dominated by simple forms such as Chuaria, Ellipsophysa, and Tawuia. But several morphological innovations did occur in the early Neoproterozoic. These include algal thalli with well-differentiated stipe and holdfast structures (in Longfengshania and Paralongfengshania), as well as cylindrical thalli with well-defined transverse annulations and holdfast structures (in Sinosabellidites and pararenicolids).

Important morphological innovations evolved in the Ediacaran Period. The Doushantuo Formation (635-550 Ma) and equivalent rocks in South China contains diverse macroalgal assemblages (Steiner, 1994; Yuan et al., 1999; Xiao et al., 2002). Doushantuo macroalgae are featured with monopodial and spiral branching (e.g., Doushantuophyton quyuani and Anomalophyton zhangzhongyingi), true dichotomous branching and apical meristematic growth (e.g., Doushantuophyton lineare, Miaohephyton bifurcatum, Konglingiphyton erecta, and Enteromorphites siniansis), rhizoidal holdfasts and flattened blade-like thalli (e.g., Baculiphyca taeniata), conical thalli (e.g., Protoconites minor), and fan-shaped thalli (e.g., Longifuniculum dissolutum, Anhuiphyton lineatum, Flabellophyton strigata, Flabellophyton lantianensis, and Huangshanophyton fluticulosum). These Doushantuo macroalgae were first reported (Zhu and Chen, 1984) from uppermost Doushantuo black shale that is less than 10 m below an ash bed dating from 551± 1 Ma (Condon et al, 2005). Subsequently, similar fossils have also been found in Doushantuo black shales in southern Anhui (Bi et al, 1988; Yan et al, 1992; Chen et al, 1994a; Yuan et al, 1999) and north-eastern Guizhou (Zhao et al., 2004). More recently, at least one member of the Miaohe biota—Enteromorphites siniansis—has also been found in the lower Doushantuo Formation in the Yangtze Gorges area (Tang et al, 2005). The lower Doushantuo Formation is estimated to be between 635 Ma and 580 Ma (Condon et al, 2005). If this estimate is correct, morphological diversification of macroalgae began after the 635 Ma Marinoan glaciation (Hoffmann et al, 2004; Condon et al, 2005) but before the 580 Ma Gaskiers glaciation (Bowring et al, 2003) and perhaps before the diversification of animals (Xiao et al, 1998b; Condon et al., 2005; Narbonne, 2005).

Despite morphological innovations in the Ediacaran, several functional forms of modern macroalgae (Littler and Littler, 1980) have not been observed in any Ediacaran assemblages. These functional forms include very thin sheet-like (leafy), calcareous, and crustose thalli, which are common in modern macroalgal flora (such as Porphyra, Ulva, and coralline red algae). The lack of leafy thalli in the fossil record may be taphonomic, but the absence of calcareous and crustose thalli in the Precambrian is probably real (Steneck, 1983). Thus, the morphological diversity of Ediacaran macroalgae, although much greater than before, may still be comparatively lower than modern macroalgae.

3.2 Quantitative Analysis: Morphospace, Body Size, and Surface/Volume Ratio

3.2.1 Methods

To quantify the morphological evolution of macroalgae in the Proterozoic, we carried out a morphospace analysis of Proterozoic macroscopic carbonaceous compressions (> 1 mm in maximum dimension, with a few exceptions) that can be reasonably interpreted as macroalgal fossils (see above). Permineralized macroalgae were not included in our quantitative analysis because of the few examples of permineralized macroalgae and because of possible preservational biases between the compression and permineralization windows. After a preliminary analysis, we also excluded in our further analysis carbonaceous compressions from the Paleoproterozoic Tuanshanzi Formation reported by Zhu and colleagues (Yan, 1995; Zhu and Chen, 1995; Yan and Liu, 1997) because at least some of these may be fragmentary microbial mats (see above) and also because their morphologies are unstable.

Table 2. List of characters (or character states) used in quantitative analysis.

Thallus morphology

Thallus differentiation

Other features

1. Spherical

9. Inferred holdfast presence

14. Transverse annulation

2. Ellipsoidal

10. Discoidal holdfast

15. Dichotomous Branching

3. Tomaculate

11. Rhizoidal holdfast

16. Monopodial or spiral branching

4. Ovoid

12. Stipe

17. Coarse branches

5. Cylindrical

13. Blade

18. Delicate branches

6. Conical

19. Colonial appearance (e.g., in some Beltanelliformis populations)

7. Fan-shaped

8. Filamentous

We collected presence/absence data of 19 morphological characters or character states of 578 carbonaceous compression specimens from 17 published monographs (Table 2 and Table 3). This literature survey was by no means exhaustive, but it included representatives of most macroalgal forms. In our analysis, all characters or character states were treated as binary presence/absence variables. We performed a non-metric multidimensional scaling (MDS) analysis of the pooled data [for a detail description of the MDS method, see (Huntley et al., 2006)]. The MDS

analysis allowed us to ordinate all specimens in a two-dimensional space (dimension 1 and dimension 2). The MDS scored specimens were then assigned to four geochronological bins (Paleoproterozoic 1800-1600 Ma; Mesoproterozoic 1600-1000 Ma; early Neoproterozoic 1000-750 Ma; Ediacaran 635-550 Ma) according to their probable age. MDS variances for dimension 1 and dimension 2 were then calculated for each geochronological bin. The sum of dimension 1 and dimension 2 variances is taken as a proxy for morphological disparity in each bin. The sum MDS variances are shown in Figs. 3-4. To test whether the geochronological pattern of MDS variance was due to varying sample intensity in the geochronological bins, we performed a randomization analysis (Huntley et al., 2006). The MDS score pairs associated with each specimen were shuffled randomly into one of the four geochronological bins, but the sample intensity of the geochronological bins was preserved. The MDS variance for each geochronological bin was recalculated. The process was repeated 1000 times, in order to obtain the mean and 95% confidence interval of the MDS

Table 3. List of geochronological bins and source data.

Paleoproterozoic (1800-1700 Ma): 29 specimens, 4 described species

Changzhougou Fm., 1800-1625 Ma

(Zhu etal., 2000)

Tuanshanzi Fm., 1800-1625 Ma

(Du and Tian, 1986)

Tuanshanzi Fm., 1800-1625 Ma

(Zhu and Chen, 1995), 10 specimens not included in further analysis

Mesoproterozoic (1600-1000 Ma): 46 specimens, 7 described species

Hongshuizhuang & Gaoyuzhuang Fm., ~1400 Ma

(Du and Tian, 1986; Walter et al, 1990)

Rohtas Fm., 1600-1000 Ma

(Kumar, 1995)

Suket Shale, 1600-1000 Ma

(Kumar, 2001)

Early Neoproterozoic (1000-750 Ma): 422 specimens, 76 described species

Liulaobei, Jiuliqiao, Jinshanzhai, Shijia, Weiji, & Gouhou Fm., ~850 Ma

(Duan, 1982; Wang and Zhang, 1984; Steiner, 1997)

Wyniatt Fm., 1077-723 Ma

(Hofmann and Rainbird, 1994)

Xiamaling, Changlongshan, & Nanfen Fm., ~850 Ma

(Duan, 1982; Du and Tian, 1986)

Halkal Formation, ~850 Ma

(Maithy and Babu, 1996)

Shihuiding Formation, ~850 Ma

(Zhang et al, 1991; Zhang et al, 1995)

Little Dal Formation

(Hofmann, 1985)

Late Riphean Pav'yuga Formation (?)

(Gnilovskaya et al, 2000)

Ediacaran (635-550 Ma): 91 specimens, 27 described species

Doushantuo Formation, 635-550 Ma

(Steiner, 1997; Xiao et al, 2002)

Lantian Formation, 635-550 Ma

(Yuan et al, 1999)

variances after randomization (Figs. 3-4). If the observed MDS variances lie beyond the 95% confidence interval, they are unlikely to be explained by differing sampling intensity alone.

To evaluate the impact of the Tuanshanzi compressions (Yan, 1995; Zhu and Chen, 1995; Yan and Liu, 1997), we repeated our analysis with the Tuanshanzi fossils included and the results did not change significantly (compare Fig. 3 and Fig. 4). As the geochronological pattern of MDS variance show no significant difference whether the Tuanshanzi compressions are included or excluded (Figs. 3-4), the Tuanshanzi fossils are excluded in all subsequent analyses (Figs. 5-8) because they are possibly fragmented microbial mats.

MDS Analysis without Tuanshanzi Material

Paleoprot. Mesoprot. early Neoprot. Ediacaran

Figure 3. Results of MDS analysis and randomization test with the Tuanshanzi material excluded. Shaded bars represent MDS variances of the four geochronological bins. Filled squares, diamonds, and triangles represent the 97.5% percentile, mean, and 2.5% percentile of the randomization test. Thus, the filled square and triangle bracket the empirically determined 95% confidence interval for each bin. Note that three of the four bins have MDS variances outside the 95% confidence intervals.

Paleoprot. Mesoprot. early Neoprot. Ediacaran

Figure 3. Results of MDS analysis and randomization test with the Tuanshanzi material excluded. Shaded bars represent MDS variances of the four geochronological bins. Filled squares, diamonds, and triangles represent the 97.5% percentile, mean, and 2.5% percentile of the randomization test. Thus, the filled square and triangle bracket the empirically determined 95% confidence interval for each bin. Note that three of the four bins have MDS variances outside the 95% confidence intervals.

MDS Analvsis with Tuanshanzi Material

MDS Analvsis with Tuanshanzi Material

Paleoprot. Mesoprot early Neoprot. Ediacaran

Figure 4. Results of MDS analysis and randomization test with the Tuanshanzi material included. See Fig. 3 for explanation.

Paleoprot. Mesoprot early Neoprot. Ediacaran

Figure 4. Results of MDS analysis and randomization test with the Tuanshanzi material included. See Fig. 3 for explanation.

* . 4

i f

A. Ediacaran _2 -

N=91

Dimension 1 (53.2%)

JA 2 * / » 1 <

*

B. early Neoprot. _2 -

—N=422

3 -I

/ 2 "

/ 1 -

1

- -

y * - 0

C. Mesoproterozoic -

N=46

Dimension 1 (53.2%)

3 i

/ 1 "

yf .. 0

a :=s

------ af

— -1

D. Paleoproterozoic _

N=19

Dimension 1 (53.2%)

1.0 n

0.5 -

k stipe discoidal holdfast f/M*

I

-1.0 -0.5

0

¡// 0.5 1.0

spherical thallus

transverse annulation

-0.5 -

tomaculate thallus

E. Character Loadings

-1.0 -

Figure 5. (A-D) Scatter plots showing realized morphospace in each geochronological bin (convex hulls in dashed line) in comparison with occupied morphospace when all Proterozoic data are pooled (convex hulls in solid line). The seemingly small occupied morphospace in Mesoproterozoic as compared with early Neoproterozoic may be related to its smaller sample size. (E) Loading diagram showing characters with significant loadings.

Scatter plots for each geochronological bin are shown in a two-dimensional space (Fig. 5A-D). Correlation coefficients were calculated between the morphological variables and MDS dimension 1 and dimension 2

scores for all species occurrences. R-values from correlation analysis were used to produce a loading chart relating the MDS morphospace to the original morphological characters (Fig. 5E).

As a proxy of body size, we also estimated the maximum dimension (e.g., long axis of an elliptical compression; maximum length of a ribbonlike compression; maximum height of a branching thallus; maximum dimension of a Longfengshania thallus including its vesicle and holdfast) of all carbonaceous compression fossils in our database. In addition, we estimated the surface/volume ratio for each specimen in our database, based on three-dimensional reconstructions of the compression fossils (see above). For example, Chuaria circularis was modelled as a spherical thallus with a diameter equivalent to its circular compression; Tawuia dalensis as a cylindrical thallus with semi-spherical ends; Longfengshania stipitata as a spherical to ovoidal vesicle with differentiated stipe and holdfast; and Doushantuophyton lineare as terete dichotomous branches with differentiated holdfast. The surface/volume ratio of Longfengshania and Paralongfengshania was estimated based on the vesicle, because it is likely that only the vesicle was photosynthetic; however, the ratio would not change significantly even if we consider the stipe and holdfast. Similarly, the holdfast of many Doushantuo macroalgae, such as Baculiphyca taeniata and Enteromorphites siniansis, was not considered in the estimate of surface/volume ratio.

3.2.2 Results

The MDS analysis (Figs. 3-5) shows that macroalgal morphospace increased episodically in the Mesoproterozoic Era and in the Ediacaran Period, confirming the narrative description. This pattern cannot be a sampling artifact because (1) MDS scores show no correlation with bin characters (data density or geochronological duration); and (2) three of the four geochronological bins have morphological disparity outside the 95% confidence interval estimated from randomization analysis (Fig. 3). In addition, a discriminant analysis shows that MDS variances of all pairwise comparisons among the four geochronological bins are significantly different (p<0.05), except the early Neoproterozoic vs. Paleoproterozoic comparison (p=0.10).

The median of the maximum dimension shows no significant change in the Proterozoic (Fig. 6). However, the range of maximum dimension expanded throughout the Proterozoic. The surface/volume ratio (Fig. 7) appears to have changed little until the Ediacaran, when both the maximum and median surface/volume ratio increased significantly (Wilcoxon test, p<0.05).

DISCUSSION

4.1 Comparison with Acritarch Morphological History

At the broadest scale, the MDS result appears to be similar to that of Proterozoic acritarchs (Huntley et al., 2006). Morphological disparity of Proterozoic acritarchs increased episodically in the early Mesoproterozoic and in the early Ediacaran, with a long-lasting plateau in between. The acritarch data also show morphospace contraction associated with Cryogenian (750-635 Ma) glaciations and late Ediacaran (575-542 Ma) radiation of Ediacara organisms. These details cannot be tested in the macroalgal data because of the poor geochronological resolution and the absence of macroalgal data in the Cryogenian and latest Ediacaran Period (550-542 Ma).

Figure 6. Maximum dimension (in mm) of Proterozoic carbonaceous compressions in linear (left) and logio scales (right). Box-and-whisker plots show median, lower and upper quartiles, and maximum and minimum values of each geochronological bin.

Given that most acritarchs were probably planktonic photoautotrophs, the first-order match between acritarch and macroalga morphological history is intriguing. The parallel between the morphological histories of Proterozoic acritarchs and macroalgae suggests an external (i.e., environmental or ecological) forcing on the morphological evolution of Proterozoic primary producers—both benthic and planktonic. Huntley et al. (2006) hypothesize that the Mesoproterozoic to early Neoproterozoic plateau of acritarch morphospace may be related to nutrient stress and a sluggish carbon cycle in approximately the same geological interval (Brasier and Lindsay, 1998; Anbar and Knoll, 2002). The macroalgal data appear to be consistent with this hypothesis, and would further imply that this environmental forcing affected both the pelagic and benthic realms. To further test this interpretation, more geochronological, chemostratigraphic (Halverson, 2006), paleoenvironmental, and paleontological data are needed to refine the temporal relationship between nutrient stress and algal evolution.

Figure 7. Surface/Volume ratio (in mm2/mm3) of Proterozoic carbonaceous compressions in linear (left) and log10 scales (right). Box-and-whisker plots show median, lower and upper quartiles, and maximum and minimum values of each geochronological bin.

Alternatively, this Mesoproterozoic—early Neoproterozoic stasis may be interpreted in ecological terms. It has been recently proposed that the radiation of late Ediacaran large acanthomorphic acritarchs, some of which are interpreted as benthos (Butterfield, 2001), was an ecological response to macrophagous grazing by early eumetazoans which, according to molecular phylogeny and molecular clock data, diverged as benthic animals between 634 and 604 Ma (Peterson and Butterfield, 2005). Using the same argument, was the morphological evolution of macroalgae held back by the absence of animal grazing in the Mesoproterozoic—early Neoproterozoic, and was subsequently accelerated by a major top-down ecological forcing in the Ediacaran when herbivorous metazoans began to evolve? One potential problem with this hypothesis is that the macroalgal morphologies that evolved in the Ediacaran (e.g., dichotomous and monopodial branching, apical meristem, rhizoidal holdfast) do not appear to be effective morphological adaptations to defend against herbivory.

Surface/Volume Ratio

The surface/volume ratio is an important physiological factor controlling the metabolic rate of modern macroalgae. Mass-specific growth rate, measured as carbon fixed per unit of body mass per unit of time, tends to be greater in macroalgal functional-form groups with higher surface/volume ratio (Littler and Littler, 1980; Littler and Arnold, 1982). This relationship remains true whether the measurements are carried out for phylogenetically related or distant macroalgae (Hanisak et al, 1988; Steneck and Dethier, 1994; Gacia et al., 1996; Stewart and Carpenter, 2003). Clearly, the effect of surface/volume ratio on macroalgal growth rate overrides phylogenetic relatedness and is pervasively convergent. Indeed, comprehensive data compilation shows that log (maximum growth rate) and log (surface/volume ratio) scale linearly over a wide range of surface/volume ratios spanning from unicellular algae, macroalgae, to rooted angiosperms (Fig. 8) (Nielsen and Sand-Jensen, 1990; Nielsen et al., 1996).

Narrative Description

Figure 8. Top: the relationship between surface/volume ratio and maximum growth rate (left vertical scale) of modern photosynthetic eukaryotes [modified from (Nielsen and Sand-Jensen, 1990)]. Mean surface/volume ratios for the four Proterozoic bins are plotted along the regression line, to show the Ediacaran increase in surface/volume ratio. Bottom: surface/volume ratio distribution (right vertical scale) of all Proterozoic macroalgae in our database.

Figure 8. Top: the relationship between surface/volume ratio and maximum growth rate (left vertical scale) of modern photosynthetic eukaryotes [modified from (Nielsen and Sand-Jensen, 1990)]. Mean surface/volume ratios for the four Proterozoic bins are plotted along the regression line, to show the Ediacaran increase in surface/volume ratio. Bottom: surface/volume ratio distribution (right vertical scale) of all Proterozoic macroalgae in our database.

The surface/volume ratios of Proterozoic macroalgae are plotted toward the lower end of modern macroalgae (Fig. 8), but did show a significant increase in the Ediacaran (Fig. 7). This pattern appears to be consistent with the complete absence of some of the extremely fast-growing functional-form groups, such as leafy macroalgae [e.g., Ulva or Porphyra; (Littler and Arnold, 1982)] in the Proterozoic.

What might have caused the Ediacaran increase in surface/volume ratio? Certainly, the greater surface/volume ratio of Ediacaran macroalgae was introduced by morphological innovations of certain functional-form groups (e.g., delicately branching forms such as Doushantuophyton, Anomalophyton, and Glomulus), which did not appear until the Ediacaran. The question is whether the Ediacaran increase in surface/volume ratio was made possible by a major evolutionary breakthrough that overcame the intrinsic developmental barriers to greater surface/volume ratios, or it was also forced by external selective pressure.

At a fundamental level, the morphogenesis of macroalgae with greater surface/volume ratio (e.g., delicately branching forms and thin leafy forms) requires parenchymatous growth and controlled cell division. The restriction of cell division to a marginal zone of meristematic cells or an apical meristem consisting of one or a few cells appears to be a key innovation in the elaboration of thallus morphology (Graham et al., 2000; Niklas, 2000). Parenchymatous and meristematic growth has been independently achieved in all three macroalgal groups—the chlorophytes, rhodophytes, and phaeophytes, suggesting that it can be achieved with relative ease. The convergent evolution of complex thalli, together with the independent diversification of Ediacaran acritarchs, points to the possible role of external forcing as part of the equation.

Algal growth requires light, nutrient, and CO2. Modern photosynthesis typically conserves <37% of the energy absorbed as photosynthetically active radiation (Falkowski and Raven, 1996), indicating that macroalgae probably have lived in light saturation even in the Paleoproterozoic when solar luminosity was about 80% of modern level (Kasting et al, 1988). Nutrient availability seems to be an unlikely driver either. Although it has been shown that nutrient uptake by micro- and macroalgae depends on surface/volume ratio (Hein et al, 1995), there is no evidence for greater nutrient availability in Mesoproterozoic oceans than in Ediacaran ones. Quite to the contrary, pelagic oceans of the Mesoproterozoic are thought to have been nutrient-limited because of the low concentration of biologically important elements such as Fe, Mo, and P (Brasier and Lindsay, 1998; Anbar and Knoll, 2002). It is possible that the coastal oceans were decoupled, in terms of nutrient availability, from the pelagic oceans in the Mesoproterozoic—a scenario that would weaken the hypothesis to invoke nutrient stress as a factor holding backing macroalgal morphological disparity in the Mesoproterozoic.

Surface-ocean CO2, on the other hand, was probably more readily available in the Mesoproterozoic Era than in the Ediacaran Period, given what we know about Proterozoic atmospheric pCO2 levels (Kaufman and Xiao, 2003). Is it possible that a drop in pCO2 level in the Cryogenian or Ediacaran Period may have forced macroalgae toward greater surface/volume ratio within their developmental possibilities to compensate for the lower pCO2 level? There is some evidence of CO2 limitation in modern macroalgae that do not use carbon concentrating mechanisms to store HCO3- as carbon source (Raven, 2003). These algae have to depend on diffusion of CO2 uptake, and their carboxylation rate is saturated at 25-35 ^M [CO2], while [CO2] in the surface ocean is only ~10 ^M (Hein and Sand-Jensen, 1997). Thus, algal growth in the absence of carbon concentrating mechanisms can be limited by [CO2] under conditions of light and nutrient saturation. Indeed, controlled experiments show that growth rate of some macroalgae increases moderately with elevated [CO2] or pCO2 levels up to 5x present atmospheric level (Gao et al, 1993; Hein and Sand-Jensen, 1997; Kubler et al, 1999). Thus, it appears that both carbon concentrating mechanisms and greater surface/volume ratios could have been physiological and morphological responses to decreasing pCO2 levels in the Ediacaran (Graham and Wilcox, 2000).

To the extent that macroalgal morphological diversification in the Ediacaran may have been driven by top-down ecological forcing by animal grazers (see 4.1), it is also possible that the Ediacaran increase in surface/volume ratio may have been caused by the same ecological process, because macroalgal surface/volume ratio may be coupled with morphological disparity. However, delicate macroalgal thalli with greater surface/volume ratio and faster growth rate (e.g., Ulva) tend to poorly defend against metazoan grazing (Littler and Littler, 1980; Steneck and Dethier, 1994), and thus would not be the predicted outcome of herbivory forcing.

Whatever the cause, greater surface/volume ratios of Ediacaran macroalgae may have had significant consequence on the global carbon cycle. Eukaryotic phytoplankton and macroalgae are important autotrophs in coastal environments where most organic carbon burial occurs in modern oceans. Thus, macroalgal bioproductivity could have considerable impact on the carbon cycle. A uniformitarian interpretation of the Proterozoic surface/volume data suggests that, on average, Ediacaran macroalgae were more than an order of magnitude more productive than those came before (Fig. 8). Did more productive Ediacaran macroalgae (and perhaps microalgae as well?) contribute to a larger dissolved organic carbon reservoir (Rothman et al., 2003), more volatile carbon cycle, and perhaps the eventual rise of oxygen level in the Ediacaran? Here again, our ability to answer these questions is limited by the poor temporal resolution of the Proterozoic geological and paleontological record.

4.3 Maximum Canopy Height

Vertically oriented benthic organisms evolved in the Mesoproterozoic or earlier. If the presence of holdfasts in some of the Tuanshanzi compression fossils is confirmed, macroalgal canopy height was already millimeters to centimeters in the Paleoproterozoic (Yan, 1995; Zhu and Chen, 1995; Yan and Liu, 1997). Bangiomorpha pubescens from the Mesoproterozoic Hunting Formation has holdfast structures and was up to 2 mm in height (Butterfield, 2000). Tawuia-like fossils from the Mesoproterozoic Suket Shale also appear to bear holdfast structures (Kumar, 2001) and they could reach up to 14 mm in height (note that scales in Fig. 8 and Fig. 11 of Kumar, 2001 were incorrect). Longfengshania stipitata from early Neoproterozoic rocks (Hofmann, 1985; Du and Tian, 1986) has well-preserved holdfasts and were centimetric in height. Early Neoproterozoic Pararenicola huaiyuanensis and Protoarenicola baiguashanensis were interpreted as possible animal fossils (Sun et al, 1986); however, new material (Fig. 2A-C) indicates that these carbonaceous compressions may represent holdfast-bearing, benthic macroalgae with a centimetric canopy height (Qian et al, 2000).

The Ediacaran Period experienced a significant expansion of macroalgal canopy height. Some of the holdfast-bearing forms from the Doushantuo Formation, such as Baculiphyca taeniata, were decimetric in height (Xiao et al., 2002). Maximum dimension of Proterozoic carbonaceous compressions, regardless whether they are benthic or planktonic, also shows a sharp increase in the Ediacaran Period (Fig. 6). Given that many specimens in our database are benthic macroalgae (with or without preserved holdfasts), the maximum dimension data can be taken as suggestive evidence that maximum canopy height was greater in the Ediacaran Period than before. The simultaneous increase in both maximum dimension and surface/volume ratio of Ediacaran macroalgae indicates greater morphological complexity, consistent with our morphometric analysis that shows a significant Ediacaran increase in MDS variance (Figs. 3-5).

4.4 Ecological Interactions with Animals

Ecological interactions among living organisms form a complex network (Fig. 9). The nature of ecological interactions includes competition, predation, symbiosis, parasitism, herbivory, and many others. Very little is known about ecological interactions in the Proterozoic ecosystem (Fig. 9). Among the few examples of ecological interactions in the Proterozoic are predation on Cloudina animals (Bengtson and Yue, 1992; Hua et al., 2003) and lichen-like algal-fungal symbiosis (Yuan et al, 2005), both are preserved in Ediacaran rocks.

Figure 9. Organismal interactions in modern ecosystems (left) compared with what we know about ecological interactions in the Proterozoic (Bengtson and Yue, 1992; Seilacher, 1999; Yuan et al., 2005). Animal-algal interactions are indirectly inferred based on arguments presented by Peterson and Butterfield (2005), not on direct fossil evidence. Modified from (Taylor et al, 2004).

Figure 9. Organismal interactions in modern ecosystems (left) compared with what we know about ecological interactions in the Proterozoic (Bengtson and Yue, 1992; Seilacher, 1999; Yuan et al., 2005). Animal-algal interactions are indirectly inferred based on arguments presented by Peterson and Butterfield (2005), not on direct fossil evidence. Modified from (Taylor et al, 2004).

Is there paleontological evidence for Proterozoic macroalga-animal interactions? Herbivory is an important form of macroalga-animal interaction, but so far we have identified no direct evidence for herbivory in Proterozoic carbonaceous compression fossils. Proterozoic macroalgal fossils, at least those >550 Ma in age, typically do not have wounds, particularly healed wounds. There might be a taphonomic issue here; after all, healed wounds are not usually preserved in the meagre fossil record of Phanerozoic macroalgae either. The complete absence of crustose calcareous algae in the Proterozoic, however, is a true signal. In fact, some phosphatized algae from the Doushantuo Formation are probably phylogenetically related to modern calcareous coralline algae but lacked biocalcification (Xiao et al., 2004). Insofar as biocalcification is an effective protection against herbivory and calcareous coralline algae depend on herbivore denudation to prevent epiphyte colonization (Steneck, 1983), the lack of calcareous algae in the Proterozoic is circumstantial evidence for the absence of herbivory, at least Phanerozoic-style herbivory; incidentally, this inference is also consistent with conclusions derived from phylogenetic arguments that herbivores appeared relatively late among animal groups (Vermeij and Lindberg, 2000). In addition, the dominance of simple discoidal and delicate rhizoidal holdfasts in Proterozoic benthic macroalgae, as well as the concurrent absence of robust holdfasts (e.g., in modern seaweeds such as Laminaria and Caulerpa), indicates that animal bioturbation in normal marine soft substrates was relatively weak and that the microbially dominated substrates were firmer and less soupy prior to ~550 Ma (Seilacher, 1999; Bottjer et al., 2000; Droser et al., 2002).

The indirect evidence for the insignificance of herbivory and bioturbations can be interpreted in three different ways: animals did not evolve until 550 Ma; they were microscopic (millimetric or smaller) prior to 550 Ma, hence leaving unrecognizable traces; or they were macroscopic but not effective herbivores or burrowers. Regardless, the limited evidence seems to suggest that macroalga-animal interactions were comparatively weak or unrecognizable in the fossil record. This is particularly true for Phanerozoic-style herbivory, but it remains to be seen whether other forms of macroalga-animal interactions, for example parasitism, commensalisms, and herbivory on microalgae, played a significant role in the ecological evolution in the Ediacaran.

Was this article helpful?

0 0

Post a comment