The downsizing of gigantic scales and large cells in the genus Mallomonas (Synurales, Chrysophyceae)

Mallomonas is the largest and most speciose genus within the Synurales, a monophyletic clade of siliceous scale-bearing organisms within the class Chrysophyceae. The genus consists of unicellular, motile, photosynthetic organisms found in freshwater localities worldwide. Mallomonas diverged from other synurophytes during the lower Cretaceous at approximately 130 Ma. Recent discoveries of fossil species were used to examine shifts in scale and cell size over geologic time. On average, scales of fossil species were 2.5 times larger than those produced by modern species. However, a smaller subset of extinct fossil taxa lacking modern analogs had scales over four times larger than modern species, and the largest recorded specimens were six times larger. Data from modern species were further used to develop a model relating scale size to cell size, and applied to the fossil specimens. Based on the model, the mean size of fossil cells was almost twice as long and 50% wider compared to modern species, and cells of taxa lacking modern analogs close to three times as large. These large cells, covered with robust siliceous scales, were likely slow swimmers requiring significant energy to maintain their position in the water column, and possibly prone to increased predation.

www.nature.com/scientificreports/ those species with such scales typically have a close-fitting and precise cell covering 3,14 . The dome is a raised cavity on the distal end of the scale into which the bristle foot, is tucked. The bristle shaft emerges from an inverted U-shaped opening along the distal margin of the dome. Scales that possess a dome and V-rib are termed tripartite scales because they have three regions, the dome, shield and posterior flange 17 .
Until the recent discovery of numerous synurophyte remains in the early Eocene Giraffe Pipe fossil locality, there was no record of this group of organisms older than Holocene 18,19 . Findings from this fossil locality have yielded preliminary insights into how scales, bristles and cells have evolved over time, and have aided in linking evolutionary events to geologic time [20][21][22] . Using relaxed molecular clock methods calibrated with Giraffe microfossils, Siver et al. 19 estimated the origin of the synurophytes to be upper Jurassic, approximately 156 Ma, with the primary genera Mallomonas and Synura diverging from each other during the lower Cretaceous at 130 Ma ( Fig. 1). A Cretaceous timeframe for divergence of Mallomonas and Synura was supported by Skaloud et al. 23 . Jo et al. 24 and Siver et al. 19 further showed that Mallomonas diverged into two major clades during the Cretaceous, one clade consisting of species with scales lacking a V-rib ( Fig. 1; subclade A2), and the other with V-rib-bearing scales ( Fig. 1; subclade A1).  19 . The mean divergence time (top number in millions of years) and percent posterior probability (bottom number) are given at each node. Clades representing Mallomonas (clade A) and Synura (clade B), along with major subclades for each genus are illustrated with representative scale types. For Mallomonas, subclades A1 and A2 include taxa that have scales with and without a V-rib, respectively. Primary parts of each Mallomonas scale type are denoted, including the dome, shield, V-rib, base plate and posterior rim (PR). Reprinted with permission from the Botanical Society of America. www.nature.com/scientificreports/ Siver et al. 19 offered preliminary findings regarding shifts in the sizes of scales since the Eocene, including identification of some fossil species bearing significantly larger scales compared to those formed by modern species. The objectives of this study are to (1) further use the fossil record to examine shifts in the size of Mallomonas scales and; (2) establish a model relating scale size to cell size, and use it to estimate the sizes of fossil Mallomonas cells.

Materials and methods
The morphometric database of scale and cell sizes used in this study was developed from one originally used by Siver et al. 19 . The initial database consisted of data for 100 modern species of Mallomonas that were taken from Siver 15 and Kristiansen 16 . Scale and cell morphometric data for an additional 24 modern species described after the Kristiansen 16 publication were taken from original descriptions and literature records, and scale data for 21fossil species (Table 1), were added to the database for this study, resulting in a total of 145 taxa (Supplementary File 1). Seventeen of the 21 fossil species have now been officially described ( Table 1). Three of the remaining four fossil species, Mallomonas GP2, M. GP4 and M. GP13, were included in Siver et al. 19 , and Mallomonas W1is illustrated for the first time in this study. Mallomonas W1 was uncovered from the 83 Ma old Cretaceous Wombat locality 25 , and the remaining fossil taxa from the 48 Ma Eocene Giraffe pipe locality 19 .
Except for data given in Siver 15 , literature records for most modern Mallomonas species include only ranges and not mean values. In lieu of the lack of mean estimates for most modern species, range midpoints for scale and cell data were used to develop models relating scale size to cell size. Range midpoint values were also used to estimate cell size for fossil species. Between 20-30, and a minimum of ten, scales were measured for each fossil species. As reported by Siver et al. 19 , the fact that there are highly significant relationships between mean and range midpoint measurements for the suite of species where both values are available, supports the use of range midpoints as a surrogate for mean values. As a result, the mean values of cell size reported in the study are equivalent to range midpoints.
Scale surface area was estimated using the formula for an ellipse: SA = (a * b * π), where a = radius of the major axis, b = radius of the minor axis, and π = 3.14. As the diameter of the minor axis approaches that of the larger axis, the formula converges on the surface area of a circle, and 78.5% of the area of a true square. Thus, it works well for estimating surface area of scales that are ellipse-shaped, and those reported as circular-shaped. Table 1. List of 21 fossil Mallomonas species used in this study. All species except Mallomonas W1, which was found in the Cretaceous Wombat locality, were uncovered from the Giraffe Pipe fossil locality. Each species is scored as a) being similar to a modern taxon; b) closely related to a modern taxon or; c) lacking a modern analog. Section assignments within the genus are according to Kristiansen 16 or as modified by Siver et al. 19 , subclade assignments after Siver et al. 19 and as shown in Fig. 1, and those species reported as stem lineages are noted. See text for details. Table modified and updated from Table 1 in Siver et al. 19 . 1 According to Jo et al. (2011) the section Heterospinae should be combined with section Planae. 2 Taxon first illustrated in this paper. 3 Taxon first illustrated in Siver et al. 19 , but not yet formally described. www.nature.com/scientificreports/ In addition, this formula provides a reasonable estimate of surface area for scales reported as "square-shaped, " since these scale types actually form rounded margins and not right-angled edges. Regression analyses were used to relate scale length, width and surface area measurements, and for exploring the relationships between scale size and both cell length and width. Relationships between any two scale size measurements were found to be highly correlated, and all three estimates of scale size were significantly related to cell size. Because cell length was most highly correlated with scale surface area, and cell width with scale length, these models were used to estimate cell length and width for all fossil species. Because of the high correlation between any combination of scale size parameters (e.g. scale length and width), adding multiple independent variables did not significantly improve either model for estimating cell size. Linear regression analyses were done using SigmaPlot v. 12.5.

Results
Scale size. The mean length and width of body scales for modern species (n = 124) ranged from 1.9-9.5 µm and 1.3-6.2 µm, and for fossil species (n = 21) from 2.5-10.5 µm and 2-7.9 µm, respectively ( Table 2). The overall mean scale size of modern species was 4.2 × 2.7 µm, yielding a mean surface area of 10.1 µm 2 . In contrast, the mean scale size of the fossil taxa was 6.6 × 4.6 µm, yielding a significantly larger scale surface area of 26.2 µm 2 ( Table 2). Fossil taxa lacking modern representatives (n = 9) had a mean scale size of 8.8 × 6.4 µm, resulting in an average scale surface area four times larger than modern species, and almost twice as large as fossil taxa with modern representatives (Table 2; Figs. 2, 3). Scale width is highly and significantly correlated with scale length, for modern taxa (r 2 = 0.88, p < 0.001), fossil taxa (r 2 = 0.81; p < 0.001) and both groups combined (r 2 = 0.88; p < 0.001; Table 3).
Of the nine extinct fossil species lacking a clear link to modern taxa, seven would be classified within the section Planae ( Fig. 1; Mallomonas subclade A2). Five of the seven have body scales that are among the largest known for the genus (Figs. 2, 3A-E), and several tend to be slightly square to rectangular-shaped (e.g. Fig. 3B-C). For example, scales of Mallomonas GP13 (Fig. 3A) are the largest recorded, with a surface area over six times larger than the average for modern Mallomonas species, and 2-2.5 × greater than the largest known modern representatives (e.gs. Fig. 3F-G). Body scales of Mallomonas Wsp1, the oldest recorded species in the genus, are distinctively square-shaped and five times larger than the average for modern species (Fig. 3B). Scales of M. schumachii (Fig. 3C) and M. media (Fig. 3D) are also massive and robust, with surface areas 4-5 times larger than the average for modern species. Although body scales of the extinct Mallomonas elephantus (Fig. 3E) had a similar surface area to those of M. bronchartiana (Fig. 3F), the modern species with the largest scales belonging to section Planae, many scales of the former taxon also possessed a large wing-like structure protruding to one side on the scale. Scales of the two remaining extinct fossil species within section Planae, Mallomonas skogstadii and M. porifera, were also large with mean values of 8.4 × 4.7 µm and 5.9 × 5.1 µm, respectively. As is the case for M. elephantus, scales of M. skogstadii have a large protruding anterior wing. Scales of M. porifera have a similar surface area to those of the modern species M. caudata (Fig. 3G), but with a distinctly circular shape. In addition to size and shape comparisons of scales of modern species in section Planae ( Fig. 3F-J), representatives of common species representing Mallomonas subclade A1 ( Fig. 1) with tripartite scales (Fig. 3K-M) and a species from section Torquatae (Fig. 3N) further emphasize the large size of the extinct fossil species.
Cell size. The mean lengths and widths of cells of modern species (n = 122) ranged from 7-60 µm, and 2.5-21 µm, with overall mean and median values of 21.4 × 10.8 µm, and 18.0 × 9.5 µm, respectively ( Table 2; Table 2. Mean, minimum and maximum estimates of scale length, width and surface area, and cell length and width. Estimates are given for modern species, all fossil species, and fossil species that lack modern congeners. *Fossil species lacking modern congeners.  The relationship between scale size and cell size of modern species was investigated for the purpose of potentially reconstructing the cell size of fossil taxa using scale size data. Both cell length and cell width were found to be significantly correlated with scale size (Table 3). Cell length was most highly correlated with scale surface area (r 2 = 0.74; p < 0.001; Fig. 4), and to a slightly lesser extent with scale length (r 2 = 0.71; p < 0.001) ( Table 3). Cell width was slightly more correlated with scale length (r 2 = 0.59; p < 0.001) than scale surface area (r 2 = 0.50; p < 0.001) ( Table 3).  www.nature.com/scientificreports/   www.nature.com/scientificreports/ The linear models relating scale surface area to cell length, and scale length to cell width, were applied to the fossil taxa in order to estimate the cell sizes of these ancient organisms ( Table 2). The estimated cell length and width of the fossil species ranged from 14.3-81.7 µm and 4.8-23.2, respectively, with a mean cell size of 39.2 × 14.9 µm (Fig. 4). The mean length of the fossil cells was almost double that of the modern species, and 50% wider (Fig. 5B2). The mean size of cells for fossil taxa that lack modern congeners was significantly larger with a mean of 59.4 × 19.6 µm, and with cells of M. GP13 reaching an estimated 81.7 × 22.7 µm (Fig. 5B4). This is in contrast to a mean cell size of 31.0 × 18.5 µm for the five most common modern species belonging to section Planae, and a cell size of 55.0 × 19.0 µm for the largest modern species in section Planae, M. bronchartiana (Fig. 5B4). Lastly, the mean size of the five smallest fossil species was 18.0 × 8.7 µm, compared to a mean of 9.3 × 5.0 µm for the five smallest modern taxa.

Discussion
Siver et al. 19 identified three categories of fossil Mallomonas species uncovered in the extensive Giraffe Pipe locality. One group of species had scales with morphological characteristics similar to, and difficult to separate from, modern congeners. Based on a morphological species concept, these could be viewed as representing the same species. A second group had morphologically different scales, but ones that could be linked to one or more modern species. The third group possessed scales that could not be directly linked to any modern species. The majority of the species contained in the latter group lacked a V-rib and well developed dome, and were considered as stem organisms within the broad section Planae. Siver et al. 19 further reported that the mean size of scales in the group containing the extinct stem taxa was larger than those fossil taxa grouped with modern congeners.
The current study adds additional modern and fossil species to the database used by Siver et al. 19 , including the oldest known taxon from the Cretaceous Wombat locality, and provides the first attempt to reconstruct cell size for fossil Mallomonas species. Based on the expanded database, several trends with respect to the evolution of scale and cell size of Mallomonas taxa can be made. First, there is a strong relationship between scale width and scale length that was similar for both fossil and modern forms. Second, as a group, fossil taxa had scales that are significantly larger than those produced by modern species, especially with respect to surface area. The five species with the largest scales belong to extinct fossil species, four of which belong to the group of stem taxa within section Planae. These scales are massive compared with modern forms, and support the concept of scale gigantism for early members of the Mallomonas clade containing species with scales that lack a V-rib and dome ( Fig. 1; subclade A2). Third, assuming the model relating scale and cell size can be applied to the geologic record, fossil species produced significantly larger cells than modern forms.
Because the models relating scale length to scale width were similar for modern and fossil species, the assumption is that the models developed relating scale size to cell size are appropriate for fossil taxa. In addition, the precise overlapping pattern of scales comprising the cell covering on modern species has recently been documented for Eocene fossil species 22 , indicating that this architectural design was well evolved by at least the early Eocene. Thus, making the assumption that other fossil taxa had similarly constructed cell coverings is reasonable, and further supports the application of the models relating scale and cell size to these fossil forms.
Based on the model estimates, the mean cell size of the fossil species is approximately twice as large as the average cell produced by modern organisms. This doubling of cell size was also observed for the smallest species. The mean size of the five smallest modern species (M. canina, M. mangofera, M. dickii, M. madagascariensis, and M. gutata) was 9.3 × 5 µm, compared to the mean cell size estimated for the five smallest fossil taxa (M. pseudohamata, M. preisigii, M. dispar, M. bakeri and M. GP4) of 18 × 8.7 µm. The cell size discrepancy is even greater for fossil species that lack modern congeners, and especially for the extinct stem species within section Planae that possessed an average cell size of 69.2 × 20.8 µm, with a maximum cell size of 81.7 × 22.7 µm for M. GP13. The scales produced by these large fossil cells were not only massive in size, but also robust and heavily silicified. It is likely that these large cells covered with large, heavy and cumbersome scales would have been slow swimmers that expended significantly more energy to maintain their position in the water column than modern species. Perhaps these cells were also more prone to predation by larger zooplankton, and a combination of decreased motility and greater predation provided the evolutionary pressure for smaller and faster cells with less dense siliceous components, and ultimately caused the demise of the large-celled fossil species. In contrast, it is also possible that the stimulus initially resulting in the evolution of the larger species was the fact that they were too big to be preyed upon by smaller invertebrates.
Several points regarding the models used to estimate cell size are warranted. First, it is important to note that because the scale sizes used to estimate cell sizes for the larger fossil taxa are at the end of the range used to produce the model, caution needs to be exercised. The assumption is being made that the linear relationship of the model holds for the larger scales, and that the linear relationship does not begin to flatten and reach a maximum cell size. However, there is no indication that the relationship is reaching an asymptote, nor reason to assume that the model would not hold for organisms that produce larger siliceous components. Second, the scale and cell size data used to produce the models consisted of the midpoint values of the ranges given in the literature. Thus, the cell sizes inferred from the models represent a midpoint estimate of the range for each species, and not an upper size limit. Third, there is more data available in the literature documenting scale size than there is for cell size for most modern Mallomonas species. Additional data on cell size, especially inclusion of mean values, may help to further fine-tune the models. Lastly, the formula of an ellipse was used to estimate scale surface area for the few species with "square-shaped" scales. Although this may slightly underestimate the surface area, using a formula for a square or rectangle would have resulted in an overestimation. Because the few species with square-shaped scales were primarily the extinct fossil taxa lacking modern congeners, their cell size may actually have been slightly larger than estimated in this study. www.nature.com/scientificreports/ Interestingly, fossil scales that have morphologically similar (identical) modern counterparts were not significantly different in size from each other, implying that their corresponding cells were also of similar size. These taxa have significantly smaller scales compared to those species with gigantic scales, and closer to the mean of modern species. Perhaps, this is why the lineages of these morphologically-identical species have survived for tens of millions of years. Despite maintaining virtually identical scale types, the degree of genetic difference from a physiological or reproductive perspective between taxa with virtually identical siliceous components remains unknown 19,23 .
The extinct scale types are not only significantly larger than those of species with modern congeners, but some have a tendency of being more rectangular to square-shaped. In contrast, fossil scale types that can be linked to modern species, along with their contemporary counterparts, tend to have elliptical-shaped scales. This is especially true of body scales 15,16,19 . Although a few smaller species of Mallomonas form spherical cells, the vast majority of species produce ellipsoidal-shaped cells, and this is especially true of species forming larger cells 15,16 . Smaller elliptical-shaped scales would be more efficient in covering a curving ellipsoidal cell surface than larger and square-shaped scales, and allow for a closer fitting cell covering. Jadrná et al. 26 recently reported that scales of the closely related synurophyte genus, Synura, have also become smaller and more elongate over geologic time, complementing the observations for Mallomonas. Taken together, these findings support the idea that the evolutionary trend for synurophyte organisms has been towards smaller, elliptical scales.
Cyanobacteria, a prokaryotic group of organisms estimated to have evolved by 3.5-3.4 Ga, represent one of the earliest known and smallest life forms on Earth 27 . Since the evolution of these early prokaryotes, Smith et al. 28 estimated that the maximum body size of subsequent life forms has increased approximately 18-fold, with large jumps occurring with the evolution of eukaryote cells, and another concurrent with the advent of multicellularity. In contrast, shifts in the sizes of siliceous scales and corresponding cells of Mallomonas species are small in comparison, within an order of magnitude, and similar to changes observed for prokaryote organisms and other unicellular protists over the Geozoic 28,29 .
Despite the overall lack of historical information on cell size for the majority of unicellular eukaryote lineages, there are data for some organisms that build resistant cell walls or coverings that are taxonomically diagnostic and become incorporated into the fossil record. Diatoms produce a siliceous cell wall known as the frustule, a structure composed of top and bottom pieces called valves that are held together with additional structures called girdle bands. Frustules, or their valve components, can be uncovered from the fossil record and used to provide a direct measure of cell size. Using this technique, Finkel et al. 29 reported that the size of planktic marine diatoms declined over the Cenozoic, and correlated the shift with abiotic forcing factors, including a rise in sea surface temperature and water column stratification. Foraminifera are heterotrophic marine protists that build shells out of calcium carbonate, the latter of which can also become part of the fossil record. Changes in the size of foraminifera shells over the Cenozoic have also been correlated with shifts in the intensity of water column stratification 30 . At this time, it is not known if the decline in cell size for Mallomonas species in the section Planae lineage recorded in the current study was the result of abiotic variables (e.g. energy expenditure or temperature), biotic factors (e.g. predation), or a combination of forcing variables.
The current study has provided a means to link scale size to cell size for Mallomonas that, in turn, can be used to trace shifts in cell size over geologic time. As additional scales of Mallomonas species are uncovered from the fossil record, the scale-to-cell size model will be a valuable tool for continuing to unravel the evolutionary history of cell size for this important photosynthetic organism. Other groups of unicellular protists, including euglyphids, heliozoids and rotosphaerids, are similar to synurophytes in that they build cell coverings using numerous overlapping siliceous scales or plates that can become fossilize. Perhaps the same technique of relating scale size to cell size could be used to develop models for these protist organisms, and similarly applied to the fossil record.
It is interesting to note that most modern Mallomonas species with large body scales are found in warm tropical regions, including M.  33 and several varieties of M. matvienkoae 16 . In addition, the modern tropical taxa M. neoampla Gusev & Siver and M. vietnamica share several rare features of their scales and bristles with fossil species recorded from the Giraffe locality, suggesting a possible link between the modern tropical and fossil floras. During the early to middle Eocene, the Earth experienced warm greenhouse conditions and lacked a cryosphere 34 . The Giraffe locality, positioned near the Arctic Circle, had an estimated mean annual temperature 17 °C warmer, and a mean annual precipitation over four times higher, than present conditions 35 . In fact, the assemblage of plants and animals in the Eocene Arctic has been described as analogous to those found today in eastern Asia 36 . Perhaps tropical regions, especially in southeastern Asia, offered refugia for some of the ancient Mallomonas lineages.
In summary, multiple extinct fossil species of the diverse and common synurophyte genus, Mallomonas, are reported here to have possessed gigantic scales that are significantly larger than those found on modern species. Based on a model relating scale to cell size, cells of fossil Mallomonas species were estimated to be, on average, twice as large as modern species. A combination of larger cells with heavy siliceous scales that fit less effectively around the cell may have resulted in slower cells more prone to predation, heavier cells requiring more energy resources to maintain their position in the water column, and ultimately their demise. Additional fossil species, especially representing other localities and time periods, will ultimately strengthen our understanding of the evolution of scale and cell size in synurophyte algae.

Data availability
Data supporting this study is given in Supplementary File 1, along with all references used to develop the dataset.