Volume 117, Issue F3
Free Access

Macroholes in stalagmites and the search for lost water

Nurit Shtober Zisu

Corresponding Author

Nurit Shtober Zisu

School of Geography and Earth Sciences, McMaster University, Hamilton, Ontario, Canada

Department of Israel Studies, University of Haifa, Mount Carmel, Haifa, Israel

Corresponding author: N. S. Zisu, School of Geography and Earth Sciences, McMaster University, Hamilton, ON L8S 4K1, Canada. ([email protected] )Search for more papers by this author
Henry P. Schwarcz

Henry P. Schwarcz

School of Geography and Earth Sciences, McMaster University, Hamilton, Ontario, Canada

Search for more papers by this author
Norman Konyer

Norman Konyer

Imaging Research Centre, St. Joseph's Healthcare, East Hamilton, Ontario, Canada

Search for more papers by this author
Tom Chow

Tom Chow

Medical Physics, Juravinski Cancer Centre, Hamilton, Ontario, Canada

Search for more papers by this author
Michael D. Noseworthy

Michael D. Noseworthy

Imaging Research Centre, St. Joseph's Healthcare, East Hamilton, Ontario, Canada

Search for more papers by this author
First published: 21 August 2012
Citations: 8

Abstract

[1] Fluid microinclusions in stalagmites have provided samples of paleowaters present during the growth of the stalagmite, but only in microliter amounts. Genty et al. (2002) discovered much larger water-filled macroinclusions in some stalagmites. Using computerized tomographic (CT) X-ray-scanning and magnetic resonance imaging (MRI) we searched for such macroinclusions in 21 stalagmites from diverse localities in North and Central America and the Caribbean. We show that most stalagmites contained numerous mm to cm-sized internal cavities (macroholes). These do not penetrate the outer surfaces which in most cases are deceptively unblemished. Some stalagmites have up to 10% average internal porosity. Two types of macroholes are distinguishable: axial holes formed during growth due to slower calcite accumulation at the axial drip site; off-axis holesformed penecontemporaneously with growth in discrete layers; these cut previous growth laminae showing that they are post-depositional. Using MRI on uncut, apparently sealed specimens, we find that very few of these cavities contain significant quantities of water although they were clearly formed while the stalagmite was being continuously bathed by drip water. Presumably, the water has escaped post-depositionally, through micro fissures, extensive connected hole system, crystal boundaries or other defects.

Key Points

  • Stalagmite's inner structure is visualized by CT and MRI
  • Water escape from macroholes
  • Axial and off-axis holes represent different genetic systems

1. Introduction

[2] Speleothems are mineral deposits formed in a cave as the result of secondary chemical precipitation from supersaturated groundwater that has entered the cave. In limestone caves, stalagmites, stalactites and flowstones mostly consist of calcite, and less commonly of aragonite or other minerals. Their fabrics, trace element composition, stable-isotope ratio and organic chemistry reflect events that occurred at the surface during their growth history [Frisia and Borsato, 2010].

[3] Most speleothems are formed when drip water outgasses carbon dioxide into the air of the cave reducing the amount of calcite that can be held in solution. The partial degassing continues until equilibrium is reached with the partial pressure of CO2 in the cave atmosphere. As a result, calcite is deposited when the supersaturated drop reaches the floor, building dome shaped layered hoods one above the other to form stalagmites [Fairchild et al., 2007; Ford and Williams, 2007].

[4] Several fabric types of calcite crystals have been recognized, such as columnar, fibrous, microcrystalline, dendritic and calcareous tufa, depending on a combination of factors such as the kinetics of growth processes, drip discharge rate, fluid supersaturation, CO2 pressure in the cave, degassing, temperature and the mechanism of transport of ions from the solution to the crystal surface [Kendall and Broughton, 1978; Frisia et al., 2000; Boch et al., 2011]. Laminar structure is also common in stalagmites, implying rhythmic alteration in deposition rate. Morphological variability is common even within an individual cave where speleothems can preserve a wide range of laminated and non-laminated forms [Railsback et al., 1994; Baker et al., 2008].

[5] Extensive research has shown the value of speleothems as isotopic recorders of past climate: Wigley [1969], Hendy [1971], Thompson et al. [1974], Yonge et al. [1985], Gascoyne [1992], Bar-Matthews et al. [1997], Frumkin et al. [1999], Fleitmann et al. [2004], Drysdale et al. [2005], McDermott et al. [2006], Mattey et al. [2008], and many others.

[6] Although some speleothems can be shown to have formed in oxygen isotopic equilibrium with their parent drip water, their temperature of formation cannot be determined without also knowing the δ18O of the drip water. Since drip water is derived from meteoric precipitation, its δ18O is dependent on temperature and geographic position (latitude, longitude and altitude), changes in the atmospheric circulation pattern, seawater composition over longer time spans, and the ice volume effect [Dansgaard, 1964; Thompson et al., 1976; Schwarcz, 1986; Lauritzen, 1995].

[7] Schwarcz et al. [1976] showed that the hydrogen isotopic composition δD of drip water could be determined by analysis of microscopic (10–300 μm) fluid inclusions (microinclusions), which constitute ∼0.1 wt% of the speleothem. Some recent studies [e.g., van Breukelen et al., 2008] have used continuous flow analyses of such inclusions from Holocene stalagmites to analyze both δ18O and δD of the included water, and shown that: a) δ18O values from the most recent deposits agree closely with contemporaneous rainfall; and b) the δ18O and δD values of older inclusions lie on the Global Meteoric Water Line (GMWL), suggesting that δ18O is not altered during storage in microinclusions. However, over the longer time periods which can be encountered in datable speleothems (up to 0.5 My), the initial δ18O is unlikely to be preserved in these microinclusions because of the large surface:volume ratio, which encourages isotopic exchange between the water and the enclosing calcite. Therefore Schwarcz et al. [1976] suggested that it is preferable to use the δD value to infer the former δ18O value, taking advantage of the fact that δD and δ18O of drip water lie on the local meteoric water line [Harmon and Schwarcz, 1981]:
urn:x-wiley:01480227:media:jgrf995:jgrf995-math-0001
where δois the so-called deuterium excess [Craig and Gordon, 1965]. Today δo = 10 ‰ (defining the GMWL).

[8] However, studies of ancient Antarctic glacial ice have shown that δo varied throughout the past, largely as a result of changes in meteorological conditions (T, relative humidity, wind speed) in the regions where water is evaporated from the sea [Jouzel and Merlivat, 1982; Jouzel et al., 2007]. Therefore, the method as originally proposed by Schwarcz et al. [1976] may result in errors in calculated paleotemperatures for speleothems older than the Holocene [McGarry et al., 2004].

[9] It would seem to be possible to resolve this problem using water trapped in macroholes. Macroholes are cavities larger than 1 mm, developed in stalagmites. Genty et al. [2002], while studying two cut and polished stalagmites from a cave in France, observed ml-sized volumes of water trapped in macroholes as fluid inclusions (macroinclusions). They manually extracted the water, and analyzed itsδD values. Based on their isotopic compositions and the ages of the enclosing speleothems, Genty et al. [2002]inferred that these were ancient samples of water trapped at the time of formation of the stalagmites. This raised the possibility that other speleothems contain macroinclusions and are filled with large enough amounts of water such that post-trappingO isotopic exchange would be negligible. It would therefore be possible to determine both δD and δ18Ofor these macroinclusions and thus to calculate the pre-Holocene deuterium excessδo [Craig and Gordon, 1965; Schwarcz et al., 1976; Harmon and Schwarcz, 1981].

[10] Doubtless, macroholes have been observed in past research, but very few studies mentioned their appearance [Lauritzen and Lundberg, 1999; Frisia et al., 2000; Mickler et al., 2004; Fairchild et al., 2007] or discussed their development [Genty, 1992; Genty and Quinif, 1996]; Frisia [1996]in her discussion of diagenesis of speleothems, attributes the development of some mm-sized holes inside speleothems to post-depositional dissolution but does not describe their distribution in detail. More attention has been given in the literature to microinclusions, which are formed at the time of speleothem growth [Kendall and Broughton, 1978; Broughton, 1983; Schwarcz and Yonge, 1983; Frumkin et al., 1999; Matthews et al., 2000; Zhang et al., 2008]. As calcite layers are deposited under constant dripping water, it is expected that this water would be trapped within the macroinclusions - just as it is trapped in microinclusions [Schwarcz et al., 1976; Serefiddin et al., 2005; Vonhof et al., 2006; van Breukelen et al., 2008]. These would therefore present attractive targets for investigation because of their potential utility in paleoclimate studies.

[11] The two common methods shown in past studies for extracting the water are based on sawing and collecting the required sample size. The methods are (1) ‘decrepitation’ – which liberates fluid inclusion water at very high temperatures (∼800°C), requires small sample sizes (∼100–150 mg) but offsets the δD values of the trapped water by −20 to −30 ‰ from those of the parent cave water, and (2) crushing under vacuum, which has a small analytical error (±3 ‰) but requires larger sample sizes (∼1 g) and thus reduces the temporal resolution [Schwarcz and Yonge, 1983; Dennis et al., 2001; Matthews et al., 2000; Fleitmann et al., 2004; Zhang, 2007; Verheyden et al., 2008; Dublyansky and Spötl, 2009].

[12] Implementation of each of these methods requires removal of small blocks of material from the stalagmite which results in damaging the initial shape and structure of the stalagmite, loss of a 3–4 mm wide calcite layer along the axial sawing plane, and contamination with recent water derived from the sawing process. Moreover, sawing and polishing the sample results in immediate leakage of water trapped within some macroholes.

[13] In the present study we aimed to identify and locate macroinclusions in various stalagmites using nondestructive techniques, to search for trapped water within macro holes, and to drill with high precision to the target hole, in order to extract water for use in determination of pre-Holoceneδo. If successful, this method would permit high temporal resolution of δo values.

1.1. Analysis of Stalagmites by CT and MRI

[14] In our attempt to identify fluid-filled macroholes, we principally restricted our study to previously uncut stalagmites. To visualize their interior structures, we used two medical imaging methods: X-ray computed tomographic scans (CT) and magnetic resonance imaging (MRI). CT imaging is well suited to detect mm to cm sized cavities inside a stalagmite but is less well suited to visualizing any water that may be contained in the cavity. As opposed to CT, MRI is very sensitive to water, but unable to detect the solid calcium carbonate structure. Combining the two imaging modalities offers the potential to do complete non-destructive mapping of the stalagmite.

[15] Magnetic resonance has been previously applied to the study of fluid flow through fractures in oil-wet fractured limestone [Fernø et al., 2007]. X-ray computed tomography was used to acquire 3D images of a stalagmite and revealed multiple growth axes within the stalagmite [Mickler et al., 2004]. These authors used the CT images to locate the most reliable sampling location for paleo-environmental isotopic analyses; their images also revealed abundant macroscopic pores.

[16] The present paper is the first study combining CT and MRI methods in the study of fluid trapped inside speleothems and in visualizing the form of holes in speleothems.

2. Methods

2.1. Samples

[17] A set of 21 stalagmites was selected from the collection at McMaster University. Most were selected on the basis of the presence of an intact cylindrical section of the original stalagmite, that is, one which had not been previously sectioned parallel to the growth axis (sagittal plane). A few of the samples were segments that had been sawed along their axis, which also permitted direct observations on the inner structure and holes. One previously unsectioned stalactite was also included. The samples represent a wide range of latitudes in North and Central America and the adjacent islands (Table 1 and Figure 1). All the samples have been subjected to CT scanning. Using images obtained from CT imaging as a guide, 12 holes larger than 3 mm were drilled, in 5 stalagmites, to confirm the observations of CT scanning and to extract possible water infill. Six stalagmites were also scanned by MRI, and based on the observed presence of water, one of these was also drilled.

Table 1. Characteristics of the Sampled Stalagmites
Numbera Location CTb MRIc Length (mm) Holes Sized (mm) Holes Typee Drilledf
81008 a Mountain Cow cave, Belize 179 Y 220 25 Ax; Ox
81008 b Mountain Cow cave, Belize 206 255 20 Ax; Ox
78031 a Waterfall c., Belize 83 100 - -
78031 b Waterfall c., Belize 94 115 8 Ox
790512 b Bermuda 116 140 10 Ox
76163 Bermuda 200 Y 220 8 Ox
77703 Crystal c., Bermuda 139 Y 160 10 Ax; Ox
77704 Crystal cave, Bermuda 218 Y 240 6 Ox 2
88808 Deep Blue cave, Bermuda 146 175 5 Ax; Ox
73027 B13 Leamington cave, Bermuda 113 Y 130 15 Ax 2
75351 Coffee River cave, Jamaica 190 Y 225 30 Ax; Ox
MRG 200407 Marengo cave, Indiana, USA 246 290 30 Ox
MRG 200407 (BMFTL) Marengo cave, Indiana, USA 205 220 10 Ox
MRG 200407 (MBFCL) Marengo cave, Indiana, USA 132 130 8 Ox
NB1 Norman-Bone cave, W. Virginia, USA 87 100 30 Ax; Ox 2
NB2 Norman-Bone cave, W. Virginia, USA 99 110 15 Ax; Ox
72009 Inner space cavern, Texas, USA 151 180 6 Ax 1
BYD 2004002 P1 (left) Vancouver Island, Canada 139 160 3 Ox
BYD 2004002 P2 (right) Vancouver Island, Canada 130 155 4 Ox
770319–1 Unknown 156 Y 175 30 Ax; Ox 2
PMC Unknown 142 Y 165 3 Ax; Ox 4
MRG 2004006 Marengo cave, Indiana, USA D.O. 200 40 Ax; Ox
67016 Norman-Bone cave, W. Virginia, USA D.O. - 30 Ax; Ox
GV1 Grapevine cave, W. Virginia, USA D.O. - 30 Ax; Ox
67020 Handline cave, W. Virginia, USA D.O. - 20 Ax; Ox
  • a Number at McMaster University collection.
  • b Number of axial plane slices at CT.
  • c Surveyed by MRI: Y = (Yes), D.O. = Direct observations.
  • d Holes size is up to xx (mm).
  • e Holes type: Axial-Ax; Off axis-Ox.
  • f Number of drill holes.
Details are in the caption following the image
North and Central America map and location of the sampled stalagmites.

2.2. CT and MRI Scans

[18] X-ray computed tomography (CT) offers a nondestructive technique for visualizing features in the interior of opaque solid objects and to obtain digital information on their 3-D geometry and topology. The most common CT scanners are medical devices, which are optimized to image biological tissues. As a result, the sensors are designed with a limited dynamic range, being able to image objects up to a density of ∼3.4 g/cm3. Also, the absolute system resolution is typically limited to ∼1 mm, even though the CT pixel (named voxel or volume element) size can be smaller.

[19] Three basic reference planes used in this paper follow anatomical reference planes: the sagittal plane which divides the body into sinistral and dextral (left and right) portions, the coronal or frontal plane divides the body into posterior and anterior portions and the axial plane or cross-section, which divides the body into upper and lower portions.

[20] The CT scanner used for this study was a GE Litespeed-RT, which was primarily designed for radiation oncology imaging. Depending on the field of view, the CT scans had pixel sizes varying from 0.2 mm to 0.5 mm. The slice spacing chosen was 1.25 mm, as we searched for target holes larger than 3 mm.

[21] CT imaging provides a complete three-dimensional image of a volume by obtaining a series of contiguous slices. Resolution in CT images is partly determined by the voxel dimensions, and partly by properties of the features being imaged [Mickler et al., 2004].

[22] We used the CT images to create a three-dimensional coordinate system in order to locate small targets inside the stalagmite body. The image can be of the complete object, or virtually sliced. Holes can be located, and their dimensions and volume can be measured. An approximate density for the interior of each hole can be determined although this proved to be ineffective for estimating whether water was present.

[23] CT also provided a reasonable method to calculate the porosity of the stalagmite, by measuring the CT numbers (named also “HU” or Hounsfield units). These are a linear transformation of the observed attenuation into relative values such that the attenuation (“radiodensity”) of pure water at STP is defined to be 1000 HU, while that of air is 0 HU. A relative porosity scale can be established relative to the X-ray absorbance determined over each feature. The scale is set to give a value of 1000 HU for a water-filled hole, 0 HU for air, and 4150 HU for pure calcite.) We divided the mean HU value by the HU value that represents the density of the calcite. Mass density (in g/cm3) is linearly related to radiodensity to a first approximation and therefore we can use the HU values to calculate porosity, as follows.

[24] P (%) = {[Dct − Dobs]/Dct} × 100 = {[HUct − HUobs]/HUct} × 100 (2) Where Dobs and Dct or HUobs and HUct are the observed density and the density of calcite respectively, expressed in g/cm3 or HU. This value includes macro as well as micro holes within the stalagmite.

[25] We subdivided the samples into concentric shells of 5 mm thickness and calculated the average density for each shell. The shells were color coded for visualization, and show the porosity differences in the center and circumference of the sample. The density was plotted against the distance from the surface of the sample.

[26] Magnetic Resonance Imaging (MRI) is a diagnostic imaging technique that combines a powerful magnetic field with computer technology to produce detailed images of tissue (organs, muscle, fat). MRI is sensitive to free or loosely bound water molecules. Water trapped inside the stalagmite should be visible to the MRI, if it exists in sufficient quantity (approximately 0.01 μl).

[27] All MRI data were collected on a 3T Signa scanner, using an 8-channel RF receiver coil designed for the human head. The stalagmite was placed on its side with the growth tip placed at the top end of the coil (superior). Images were collected in the coronal plane. When introduced to the magnetic field of the MRI, the hydrogen atoms in the water tend to align with the magnetic field, proportionally to the strength of the MRI signal. The distribution of hydrogen spins is governed by the Boltzmann distribution, which is inversely proportional to the temperature of the water. For this reason, the stalagmite was cooled to 4°C prior to imaging in order to enhance the signal from the trapped water. Several MRI pulse sequences were explored to determine the method best suited to visualize the trapped water. The following two pulse sequences were used:

[28] 1) Spin-Echo, repetition time (TR) = 1000 ms, echo time (TE) = 13 ms, slice thickness = 3 mm, 29 contiguous slices, field-of-view = 28 cm, 256 × 256 acquisition matrix, 4 averages, scan time = 17 min and 2) Free Induction Decay Chemical Shift Imaging (fidCSI), TR = 2000 ms, TE = 0 ms, non-slice selective RF pulse (entire stalagmite projected into single image), FoV = 28 cm, 256 × 256 acquisition matrix, 2 averages, scan time = 17 min.

[29] Scanning by MRI presents the problem that the calcite material of the sample is not revealed because MRI responds only to protons or other high-spin nuclei. Therefore, in order to determine the coordinates of the location where water was located inside the sample, we used three techniques to generate an image of its external form: 1) tablets of vitamin E (which has a strong MRI signal) were attached to several widely spaced points on the surface of the sample, 2) the sample was wrapped with water-soaked cloth to make its outer form clearly visible, 3) the sample was wrapped with 3 mm diameter plastic tubing filled with a 3 mg/ml solution of a gadolinium compound called Omniscan (gadodiamide) (Figure 2) which has an intense signal in MRI . The first two methods were used in preliminary scans and the last was only used on one stalagmite that showed water filled inclusions under the first two preliminary scans.

Details are in the caption following the image
Sample preparation for MRI: Sample 77704 has been wrapped in a net of tubes, filled with a solution of the Gd contrast agent Omniscan, in order to mark the stalagmite contours. The tube content is visible on MRI scan and the plastic tubes are visible on CT scan (see also Figure 11).

[30] Where holes were detected in the CT images, we attempted to recover fluid (water) trapped in these holes by drilling into them. For this purpose, the samples were covered with plastic wrap, and placed in a rectangular box. Polyurethane foam was injected to fill the empty space, to stabilize the stalagmite in place, and to define a vertical angle which allowed precise orientation of the drill (Figure 3). The stalagmites were drilled with a 1.8 mm diameter bit. The location of the hole had been defined in relation to CT-opaque markers attached to the surface of the stalagmite. In order to avoid possible contamination with calcite powder created by the drilling process, we drilled slowly up to a distance of 1–2 mm from the target and broke the remaining barrier with a fine chisel. The powder was blown out with compressed air at each mm or less drilled. In most cases we subsequently re-examined the sample in CT to confirm that we had penetrated the hole. The same drilling method was used to test holes identified by MRI.

Details are in the caption following the image
Stalagmite sample embedded in polyurethane foam and mounted for drilling on platform of drill press.

3. Results

[31] In Table 1we give the geographic location of each sample, and some of its physical characteristics. The ages of only a few of these samples have been determined by U-series dating but this information is not presented here as it is irrelevant to the issue of identifying water-filled holes. Where CT scanning revealed the presence of holes in the interior of a stalagmite, the maximum size of these holes is shown inTable 1. Note that holes 3 mm in diameter or larger are present in all but four of the stalagmites, suggesting that internal macroporosity is a widespread phenomenon in stalagmites.

[32] The external form of a stalagmite from Norman Bone cave [Thompson et al., 1976] gave no clue that it was riddled with an extensive system of internal cavities (Figure 4). These comprise a significant fraction (10%) of the volume of the stalagmite, and most holes are located along the stalagmite core but off-axis. InFigure 4b we see an individual, elongated hole (green) as it would have projected into the adjacent mass of stalagmite (not shown in the image). The CT images allow us to define the shape and size of these holes with high spatial resolution (∼1 mm).

Details are in the caption following the image
CT images of a 10 cm long segment of stalagmite NB1 from Norman Bone cave, West Virginia. (a) The stalagmite in relation to its coordinate system. (b) Visualization of a hole (green) extended into the space occupied by the adjoining slice. (c) Visualization of the holes, without the surrounding calcite. (d) Large and well-connected holes (red) visualized in coronal plane. These make up about 10% of the volume of the stalagmite. Note the beams marking projected drill paths.

[33] In general, we observed a wide range of internal porosity in stalagmites, characterized by varying spatial distribution, size, and forms. Stalagmites vary from compact (few holes) to “spongy” in overall internal structure. Holes are located both along and off the axis of the stalagmite, commonly along the stalagmite core. An outer layer, a few mm thick in each stalagmite, is usually free of holes (shown as a distinctive color in Figure 4).

[34] In order to evaluate the calcite porosity, the stalagmites have been subdivided into concentric shells of 5 mm thickness (Figure 5b). NBC represents the most porous stalagmite (Figure 5a). Its large holes are represented by a very porous core: 17–18% porosity along the axis, while its “skin” is formed of very dense calcite (0% porosity). PMC represents a less porous stalagmite, as its maximum porosity values are less than 4%. A stalagmite of intermediate porosity is represented by 78031b, with maximum porosity values of 8%. Most holes are developed away from the growth axis, as the porosity of the inner shell decreases to 5.8%.

Details are in the caption following the image
Porosity measurements. (a) Three representative stalagmites showing minimum, maximum and average porosity values. (b) Concentric porosity shells constructed for NBC stalagmite. Serial shells correspond to points on the graph for this stalagmite.

[35] The inner structures of stalagmites are very diverse (Figure 6). A few stalagmites were almost hole-free, and CT images displayed only clear growth layers (Figure 6c). Figure 6ashows a generally hole-free stalagmite whose growth direction changed several times through its history. A single axial hole is visible; note that the growth layers dip into the hole, suggesting that the active growth surface of the stalagmite was dimpled by this hole which eventually was sealed by subsequent layers. Other sections parallel to this one demonstrate that the axial hole does not extend further along the growth axis.

Details are in the caption following the image
Appearance and distribution of internal holes in various stalagmites and a stalactite, on CT imaging; all images are displayed in sagittal (vertical) planes; arrows on panel A and C mark changes in the growth direction of the stalagmite axis; the gray scales correspond to relative X-ray attenuation, as a function of elemental composition and density, with brighter tones representing higher HU values. Note the denser skin in comparison with the inner core. The gray scale varies from panel to panel in order to generate the best image. The HU values for each panel are as follows [minimum (darkest, excluding black = air) - maximum (white)]: (a) Bermuda 73027, Leamington cave, 13 cm long: 2900–4500; (b) NB1 form Norman Bone cave, West Virginia, 10 cm long: 3400–4200; (c) Belize 78031A, Waterfall cave, 10 cm long: 3400–4200; (d) PMC, unknown location, 16.5 cm long, drilled, 3800–4300; (e) Bermuda 77704, Crystal cave, 24 cm long: 3800–4200 ; (f) Bermuda 73020, Quarry cave, 22 cm long–Stalactite: 3800–4200.

[36] Other stalagmites are characterized by a very porous structure in which holes can follow a particular growth layer, or may be dispersed through the whole body. Individual holes range up to 3 cm in maximum dimension, and are invariably elongated nearly parallel to the axis of the stalagmite. Some contiguous holes appear to be joined by openings in their mutual walls. Figure 6b (NBC, also shown in Figures 4 and 5) displays a complex network of holes that are mostly distant from the growth axis and penetrate the growth layers. Note that the layers do not change their direction at the edge of the hole, indicating that the hole formed post-depositionally, cutting previous growth layers.

[37] The stalagmite in Figure 6d, contains several mm-size off-axis holes and was drilled to search for water; the two drilled holes are visible on the left side of the figure.Figure 6e shows a stalagmite, in which the calcite crystals grew radially from a basal nucleus. Numerous holes, mainly off axis, tend to follow the crystal orientation.

[38] Stalactites have a completely different inner structure than stalagmites, as is well known from previous works [Kendall and Broughton, 1978; Broughton, 1983; Ford and Williams, 2007] and clearly visible in the lone stalactite that was scanned. A long void is developed along the axis, and calcite layers are growing parallel to the main void, in an outward direction. Small off-axis holes were observed, preferentially distributed within certain layers, precipitated concentrical to the axis (Figure 6f).

[39] Further aspects of holes in stalagmites are also visible in cross-sectional CT views (Figure 7). Figure 7ashows a typical axial hole with two small off-axis holes. Holes developed peripheral to the main axis are shown inFigures 7b and 7c. In Figure 7c some holes appear to be separated by thin septa or connected through their walls.

Details are in the caption following the image
CT imaging of axial (cross-sectional) planes of three types of stalagmites. (a) Bermuda 73027b: axial type; (b) Bermuda 76163: off axis type; (c) NB1, W. Virginia: spongy type. Note drill hole used to test for presence of water.

[40] Overall, these observations reveal that most stalagmites (and one stalactite) contain macroholes which do not extend to their outer surfaces. These features are clearly different in appearance from fluid inclusions previously described in speleothems [Schwarcz et al., 1976; Kendall and Broughton, 1978; Scheidegger et al., 2010], especially with respect to their size. The macroholes which we have observed appear to resemble in size and form those previously described by Genty et al. [2002].

[41] Despite the clear visualization of holes in the interior of stalagmites, conventional diagnostic CT imaging is not effective at determining whether the holes are filled with water or with air. The values obtained for some larger holes were approximately 500–600 HU, indicating that the voxels included some calcite crystals (Figure 8) as well as empty (air filled) spaces between them; the value was distinctly less than the value for water. Another example was a 13 mm long hole in stalagmite “Bermuda 73027” (Figure 6a). It displayed a minimum value of ∼450 HU at its center increasing to the HU value of calcite (4125) near the edge of the hole. This suggests that calcite crystals had grown toward the middle of the void.

Details are in the caption following the image
Four cross sections in three of the surveyed stalagmites, demonstrating HU values in the CT images. The scale is set to give a value of 1000 HU for a water-filled hole, 0 HU for air, and 3000–4000 HU for calcite, depending on its chemical composition. The values obtained for some larger holes were approximately 500–600 HU, indicating possible mixture of air and calcite.

[42] A total of 12 holes were drilled in 5 stalagmites, but all of them were found to be empty, even the PMC stalagmite, whose holes appeared in CT to be disconnected and sealed (Figure 9).

Details are in the caption following the image
Sample PMC visualized in 3-D by CT scanning. Two of the holes are visible in the three planes. The holes were drilled with high precision, but no water was found.

[43] MRI images of seven stalagmites with large holes identified by CT, likewise showed the complete absence of water. Figure 10shows an MRI image of two stalagmites, the outlines of which are defined by the image of water-soaked cloths that had been wrapped around them. Although CT scans of each showed abundant macroholes, they all appeared to be devoid of water.

Details are in the caption following the image
Water artificially introduced into stalagmites: In order to delimit their margins, the stalagmites have been wrapped in a wet cloth. (a) Sample 770319–1 (top of stalagmite to right) has been immersed in water, allowing water to enter the previously visualized pore system; white dots show the inserted water; right-hand image is CT scan of same sample showing numerous macroholes; (b) Stalagmite PMC (top of stalagmite to left) had been drilled to check for water; a water-filled tube has been inserted into the drill hole; Vitamin E capsules provides a reference point on the sample. Note: it is difficult to obtain exact co-alignment of the magnetic field of the MRI with the frame of reference of the CT scanner; the images are intended to show the general appearance of equivalent volumes using the two techniques.

[44] It was considered possible that some characteristics of stalagmitic calcite might make it impossible to detect water by MRI even if present. To test this we immersed one of these samples (770319–1) in water for one hour, allowing water to penetrate along the cut axis of the stalagmite. The MRI image of the previously water-free sample now showed droplets widely distributed through it (Figure 10a), presumably residing in the open pore system of this sample shown in CT scans. We also inserted a water-filled tube into the hole that had been drilled into another sample (PMC).Figure 10b shows that the water filled tube was readily detectible inside the stalagmite.

[45] Stalagmite 77704 from Crystal cave in Bermuda was the only stalagmite in which clear traces of water were visible in MRI images. This stalagmite is composed of, large (2–4 cm) crystals of clear, transparent calcite resembling in this respect the stalagmites studied by Genty et al. [2002]. As observed in the MR images, the water location did not correspond exactly to the cavities found in the CT images (Figure 11). This discrepancy may be partly explained by inhomogeneities in the magnetic field of the MR instrument, which are normally partially corrected by varying (“shimming”) the magnetic field. This was not possible with the stalagmite due to the very small water signal detected.

Details are in the caption following the image
Stalagmite 77704. (a) CT image with MRI overlay. Water appears as red dots. The red ellipse shows the target inclusions. The white ellipses indicate the vitamin E markers. Note the drill hole, at a distance of 2 mm adjacent to the target inclusions. This hole was eventually extended to reach the hole but no water was recovered. (b) MRI image of stalagmite 77704. The external white dots indicate the Omniscan + water filled tubes placed as markers to define the calcite margins (see also Figure 2). The encircled dots mark the target + inclusions.

[46] Sites in 77704 where water was detectible by MRI were drilled, but no significant amount of water was recovered.

4. Discussion

[47] Occasionally, stalagmites display a deceptively dense skin, covering a rather vuggy core. Their external shape and layers do not in most cases hint at the inner structure, which remained undetectible until now, unless the stalagmite was sawed and polished. Using samples selected at random from the collection at McMaster University, representing a wide range of environments, we observe that most stalagmites contain a significant proportion of open space (not water filled) in the form of macroholes up to several cm in maximum dimension. Some larger holes, up to 3 cm, are connected to one another, either directly or by fissures and passages.

[48] We have shown for the first time the full 3-D distribution of these macroholes and that CT scanning of stalagmites is a useful nondestructive tool for imaging the interior of spelothems, revealing their 3-D geometry and topology, without damaging or losing valuable material by sawing or breakage. While CT imaging could not demonstrate whether these holes were water or air-filled, we were able to test this using drilling and MRI imaging. The advantage of the MRI procedure is that very small amounts of water can be recognized, in the range of a few microliters. The disadvantage is that the air and the surrounding calcite lack a magnetic signal and therefore the empty holes and the stalagmite material are invisible on MR. Only one stalagmite (# 77704) appeared to contain any water, but subsequent drilling of this single water-bearing sample failed to yield any free liquid. Presumably the amount was sufficient to be detectible by MRI but insufficient to permit extracting a sample of the water. The walls of the holes when seen in sectioned stalagmites appear to be rough and partly overgrown with calcite crystals. It is likely that what little water was present in this stalagmite was held by surface tension between surface features inside the holes.

[49] If the holes which are visualized by CT were formed during the growth of the stalagmites, then at the time of their formation they should have been at least partly filled with water. Microinclusions are always water-filled although small air bubbles may also be present [Scheidegger et al., 2010], but at present most macroinclusions are filled only with air. Our initial suspicion for this discrepancy was that microfractures had developed as a result of freezing and thawing during glacial-interglacial transitions as would have been experienced by pre-Holocene stalagmites. However, at least four of our MRI tested samples grew in warmer regions where periglacial conditions had never been experienced. For example, the Bermuda sample (77704, recovered from undersea and therefore dating to a glacial period) would never have experienced sub-zero temperatures and yet contained only traces of water in its holes. It is possible that water has leaked out of the macroholes through a network of permeable openings in the stalagmite of unknown origin, cracks, micro fissures, crystal boundaries or other small defects, formed either initially in the cave (by minor movements, earthquakes, etc.), or since their collection.

[50] On the contrary, previous studies showed that water is tightly stored in microinclusions of most stalagmites. These occur between and inside the crystals, and make up c. 0.1–1% of the total volume of the stalagmite. These features are certainly syngenetic as shown by isotopic analyses [e.g., Zhang, 2007] and belong to a different genetic category than the macroholes reported on here.

4.1. Form and Distribution of Macroholes

[51] We observed two types of macroholes, according to their location and relationship to the growth layers: axial holes and off axis holes (Figures 12 and 13).

Details are in the caption following the image
Cut and polished sections of stalagmites: (a) stalagmite 81008:2 showing holes formed along the growth axis (dashed line) within a sunken central splash cup: growth layers consequently dip slightly into the hole. Off-axis holes (marked by red arrows) crosscut growth layers which are undisturbed in the neighborhood of the hole, implying post-depositional development; (b) stalagmite MRG2004006: detail of off-axis holes (black arrows) showing more clearly the lack of disturbance of growth lines by presence of holes. Bubbles in the image are a result of wetting sample with water in order to visualize growth layers inside holes.
Details are in the caption following the image
CT image of stalagmite NB1. Three planes are crossing one of the large off-axis holes, marked by red cross. The coronal plane is sliced 10 mm off-axis. Stalagmite axis is visible only in the axial plane figure. All holes in the Coronal plane are off-axis holes.

[52] Axial Holes.Growth layers (visualized in CT) are observed to dip into the axial holes, suggesting that the surface of the hole was part of the active growing surface of the stalagmite. Its indentation into the drip-site on the top of the stalagmite is presumably due to the fact that the rate of calcite deposition was lower at this point; This could have been simply because deposition rate increases as the drip water progressively loses CO2 to the cave atmosphere, a process that has just started where the falling drop first meets the stalagmite surface [Romanov et al., 2008]. Spatially, axial holes are obviously concentrated along the growth axis but otherwise do not seem to be restricted to distinctive layers of stalagmites. Rather, they appear to be developed over long periods of calcite deposition. Where exposed in cross section, their inner surfaces are typically smooth or lightly covered with secondary sparry overgrowths.

[53] Off-Axis Holes. These holes invariably crosscut growth layers (Figures 11 and 12b). The holes are elongated parallel to the growth axis and approximately ellipsoidal in overall form. They tend to be more abundant in specific growth zones (series of layers) (Figures 6b, 6d, and 6e). Their inner surfaces are commonly covered with extensive sparry calcite overgrowth. Where it is exposed, the underlying wall of the hole is irregular with many indentations a few tens to 100 s of μm deep. The fact that the off-axis holes crosscut growth-layers clearly shows that these holes were formed post-depositionally, that is, by internal erosion of the previously formed stalagmite. The holes are today invisible from the surface and therefore not open for macroscopic leakage. The fact that these holes are to some extent concentrated in specific growth layers could have at least two possible interpretations: a) the holes were formed penecontemporaneously with the growth layer but after cessation of deposition of the layer which is cut by the hole; the hole ceased to form after some change in growth conditions had occurred; or b) particular growth layers are especially susceptible to hole formation, although the hole-forming process occurred long after speleothem growth.

[54] Although axial and off axis holes have been observed in past research, they have never been studied as particular geomorphic phenomena. In the present study, a differentiation from the general terms “calcite macroporosity” or “calcareous tufa fabric” [Frisia et al., 2000; Frisia and Borsato, 2010; Boch et al., 2011] is proposed for the first time, implying that some macroholes have developed after the deposition of the “hosting” bands, and therefore, they are possibly represented by a different evolutionary history. The origin of the off-axis holes is as yet unknown, and will be subjected to further observations and research.

5. Conclusions

[55] 1. CT imaging cannot demonstrate whether macroholes in stalagmites are filled with water or air. In contrast, the MRI procedure can recognize very small amounts of water, but cannot distinguish between calcite and air. Therefore, the best procedure for locating and recognize the hole content of a given stalagmite is by combining both methods.

[56] 2. Water was found in only one of the 22 stalagmites investigated, and only in small amounts. Further search for trapped water should concentrate on stalagmites composed of large, clear, crystals like that from Bermuda (77704) or the stalagmites studied by Genty et al. [2002]. However, since we have shown that water can apparently escape from macroholes, it is also possible that the small amounts found in some holes entered long after the holes were formed, through micro fissures (Figures 12 and 13).

[57] 3. Analysis of the form and distribution of the macroholes shows that axial holes are formed during growth of the stalagmite, whereas off-axis holes are clearly post-depositional. Since the off-axis holes were presumably formed by dissolution of their host speleothems and are entirely enclosed inside them, it seems likely that they were initially full of water. This water was subsequently lost by leakage along an interconnected system of crystal boundaries, or through micro fissures or other defects. We note that the axial holes which surely must also have initially contained water, are also found to be empty. Therefore, loss of water from macroholes in seemingly intact, uncut speleothems is a widespread phenomenon which is itself worthy of future study.

Acknowledgments

[58] We thank the staffs at the St. Joseph Health Care Centre and the Juravinski Cancer Centre for their assistance in generating the images presented and discussed here. We are thankful to Martin Knyf who helped in the lab, Derek Ford for his keen interest and helpful discussions and Amos Frumkin for his beneficial comments. Patricia Beddows assisted in the photography of a cut stalagmite. This research was funded by a grant from the Natural Sciences and Engineering Research Council of Canada. N.S. thanks the University of Haifa for funding her post-doctoral fellowship to McMaster University.